fix(chat): repair inconsistent quote escaping in edit_diagram JSON

When the LLM generates edit_diagram tool calls, it sometimes produces
inconsistent quote escaping in XML attributes within JSON strings.
For example: y="-20\" instead of y=\"-20\"

This causes JSON parsing to fail, and jsonrepair cannot fix this pattern.

Added pre-processing regex to detect and fix cases where the opening
quote is unescaped but the closing quote is escaped in attribute values.

.github/renovate.json

@@ -1,7 +1,7 @@
 {
     "$schema": "https://docs.renovatebot.com/renovate-schema.json",
     "extends": ["config:recommended"],
-    "schedule": ["after 10am on saturday"],
+    "schedule": ["after 10am on the first day of the month"],
     "timezone": "Asia/Tokyo",
     "packageRules": [
         {
@@ -13,6 +13,7 @@
         {
             "matchUpdateTypes": ["major"],
             "matchPackagePatterns": ["*"],
+            "groupName": "major dependencies",
             "automerge": false
         },
         {

.github/workflows/electron-release.yml

@@ -11,7 +11,8 @@ on:
         required: false
 
 jobs:
-  build:
+  # Mac and Linux: Build and publish directly (no signing needed)
+  build-mac-linux:
     permissions:
       contents: write
     strategy:
@@ -20,13 +21,9 @@ jobs:
         include:
           - os: macos-latest
             platform: mac
-          - os: windows-latest
-            platform: win
           - os: ubuntu-latest
             platform: linux
-
     runs-on: ${{ matrix.os }}
-
     steps:
       - name: Checkout code
         uses: actions/checkout@v6
@@ -40,7 +37,58 @@ jobs:
       - name: Install dependencies
         run: npm install
 
-      - name: Build and publish Electron app
+      - name: Build and publish
         run: npm run dist:${{ matrix.platform }}
         env:
           GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
+  # Windows: Build, sign with SignPath, then publish
+  build-windows:
+    permissions:
+      contents: write
+    runs-on: windows-latest
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v6
+
+      - name: Setup Node.js
+        uses: actions/setup-node@v6
+        with:
+          node-version: 24
+          cache: "npm"
+
+      - name: Install dependencies
+        run: npm install
+
+      # Build WITHOUT publishing
+      - name: Build Windows app
+        run: npm run dist:win:build
+        env:
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Upload unsigned artifacts for signing
+        uses: actions/upload-artifact@v6
+        id: upload-unsigned
+        with:
+          name: windows-unsigned
+          path: release/*.exe
+          retention-days: 1
+
+      - name: Sign with SignPath
+        uses: signpath/github-action-submit-signing-request@v2
+        with:
+          api-token: ${{ secrets.SIGNPATH_API_TOKEN }}
+          organization-id: '880a211d-2cd3-4e7b-8d04-3d1f8eb39df5'
+          project-slug: 'next-ai-draw-io'
+          signing-policy-slug: 'test-signing'
+          artifact-configuration-slug: 'windows-exe'
+          github-artifact-id: ${{ steps.upload-unsigned.outputs.artifact-id }}
+          wait-for-completion: true
+          output-artifact-directory: release-signed
+
+      - name: Upload signed artifacts to release
+        uses: softprops/action-gh-release@v2
+        with:
+          files: release-signed/*.exe
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/test.yml

@@ -11,10 +11,10 @@ jobs:
     name: Lint & Unit Tests
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v6
 
       - name: Setup Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v6
         with:
           node-version: "20"
           cache: "npm"
@@ -32,10 +32,10 @@ jobs:
     name: E2E Tests
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v6
 
       - name: Setup Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v6
         with:
           node-version: "20"
           cache: "npm"
@@ -44,7 +44,7 @@ jobs:
         run: npm ci
 
       - name: Cache Playwright browsers
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         id: playwright-cache
         with:
           path: ~/.cache/ms-playwright
@@ -67,7 +67,7 @@ jobs:
           CI: true
 
       - name: Upload test results
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v6
         if: always()
         with:
           name: playwright-report

.gitignore

@@ -67,4 +67,6 @@ CLAUDE.md
 .spec-workflow
 
 # edgeone
-.edgeone
+.edgeone
+opencode.json
+ai-models.json

Dockerfile

@@ -30,6 +30,10 @@ ENV NEXT_PUBLIC_DRAWIO_BASE_URL=${NEXT_PUBLIC_DRAWIO_BASE_URL}
 ARG NEXT_PUBLIC_SHOW_ABOUT_AND_NOTICE=false
 ENV NEXT_PUBLIC_SHOW_ABOUT_AND_NOTICE=${NEXT_PUBLIC_SHOW_ABOUT_AND_NOTICE}
 
+# Build-time argument for subdirectory deployment (e.g., /nextaidrawio)
+ARG NEXT_PUBLIC_BASE_PATH=""
+ENV NEXT_PUBLIC_BASE_PATH=${NEXT_PUBLIC_BASE_PATH}
+
 # Build Next.js application (standalone mode)
 RUN npm run build
 

README.md

@@ -40,11 +40,12 @@ https://github.com/user-attachments/assets/9d60a3e8-4a1c-4b5e-acbb-26af2d3eabd1
     - [Installation](#installation)
   - [Deployment](#deployment)
     - [Deploy to EdgeOne Pages](#deploy-to-edgeone-pages)
-    - [Deploy on Vercel (Recommended)](#deploy-on-vercel-recommended)
+    - [Deploy on Vercel](#deploy-on-vercel)
     - [Deploy on Cloudflare Workers](#deploy-on-cloudflare-workers)
   - [Multi-Provider Support](#multi-provider-support)
   - [How It Works](#how-it-works)
   - [Support \& Contact](#support--contact)
+  - [FAQ](#faq)
   - [Star History](#star-history)
 
 ## Examples
@@ -185,7 +186,7 @@ Check out the [Tencent EdgeOne Pages documentation](https://pages.edgeone.ai/doc
 
 Additionally, deploying through Tencent EdgeOne Pages will also grant you a [daily free quota for DeepSeek models](https://pages.edgeone.ai/document/edge-ai).
 
-### Deploy on Vercel (Recommended)
+### Deploy on Vercel 
 
 [![Deploy with Vercel](https://vercel.com/button)](https://vercel.com/new/clone?repository-url=https%3A%2F%2Fgithub.com%2FDayuanJiang%2Fnext-ai-draw-io)
 
@@ -206,11 +207,13 @@ See the [Next.js deployment documentation](https://nextjs.org/docs/app/building-
 -   OpenAI
 -   Anthropic
 -   Google AI
+-   Google Vertex AI
 -   Azure OpenAI
 -   Ollama
 -   OpenRouter
 -   DeepSeek
 -   SiliconFlow
+-   ModelScope
 -   SGLang
 -   Vercel AI Gateway
 
@@ -219,6 +222,10 @@ All providers except AWS Bedrock and OpenRouter support custom endpoints.
 
 📖 **[Detailed Provider Configuration Guide](./docs/en/ai-providers.md)** - See setup instructions for each provider.
 
+### Server-Side Multi-Model Configuration
+
+Administrators can configure multiple server-side models that are available to all users without requiring personal API keys. Configure via `AI_MODELS_CONFIG` environment variable (JSON string) or `ai-models.json` file.
+
 **Model Requirements**: This task requires strong model capabilities for generating long-form text with strict formatting constraints (draw.io XML). Recommended models include Claude Sonnet 4.5, GPT-5.1, Gemini 3 Pro, and DeepSeek V3.2/R1.
 
 Note that the `claude` series has been trained on draw.io diagrams with cloud architecture logos like AWS, Azure, GCP. So if you want to create cloud architecture diagrams, this is the best choice.
@@ -245,6 +252,10 @@ For support or inquiries, please open an issue on the GitHub repository or conta
 
 -   Email: me[at]jiang.jp
 
+## FAQ
+
+See [FAQ](./docs/en/FAQ.md) for common issues and solutions.
+
 ## Star History
 
 [![Star History Chart](https://api.star-history.com/svg?repos=DayuanJiang/next-ai-draw-io&type=date&legend=top-left)](https://www.star-history.com/#DayuanJiang/next-ai-draw-io&type=date&legend=top-left)

app/[lang]/about/cn/page.tsx

@@ -10,18 +10,7 @@ export const metadata: Metadata = {
     keywords: ["AI图表", "draw.io", "AWS架构", "GCP图表", "Azure图表", "LLM"],
 }
 
-function formatNumber(num: number): string {
-    if (num >= 1000) {
-        return `${num / 1000}k`
-    }
-    return num.toString()
-}
-
 export default function AboutCN() {
-    const dailyRequestLimit = Number(process.env.DAILY_REQUEST_LIMIT) || 20
-    const dailyTokenLimit = Number(process.env.DAILY_TOKEN_LIMIT) || 500000
-    const tpmLimit = Number(process.env.TPM_LIMIT) || 50000
-
     return (
         <div className="min-h-screen bg-gray-50">
             {/* Navigation */}
@@ -108,42 +97,6 @@ export default function AboutCN() {
                                 </p>
                             </div>
 
-                            {/* Usage Limits */}
-                            <p className="text-sm text-gray-600 mb-3">
-                                当前使用限制：
-                            </p>
-                            <div className="grid grid-cols-3 gap-3 mb-5">
-                                <div className="text-center p-3 bg-white/60 rounded-lg">
-                                    <p className="text-lg font-bold text-amber-600">
-                                        {formatNumber(dailyRequestLimit)}
-                                    </p>
-                                    <p className="text-xs text-gray-500">
-                                        请求/天
-                                    </p>
-                                </div>
-                                <div className="text-center p-3 bg-white/60 rounded-lg">
-                                    <p className="text-lg font-bold text-amber-600">
-                                        {formatNumber(dailyTokenLimit)}
-                                    </p>
-                                    <p className="text-xs text-gray-500">
-                                        Token/天
-                                    </p>
-                                </div>
-                                <div className="text-center p-3 bg-white/60 rounded-lg">
-                                    <p className="text-lg font-bold text-amber-600">
-                                        {formatNumber(tpmLimit)}
-                                    </p>
-                                    <p className="text-xs text-gray-500">
-                                        Token/分钟
-                                    </p>
-                                </div>
-                            </div>
-
-                            {/* Divider */}
-                            <div className="flex items-center gap-3 my-5">
-                                <div className="flex-1 h-px bg-gradient-to-r from-transparent via-amber-300 to-transparent" />
-                            </div>
-
                             {/* Bring Your Own Key */}
                             <div className="text-center">
                                 <h4 className="text-base font-bold text-gray-900 mb-2">
@@ -339,11 +292,13 @@ export default function AboutCN() {
                         </li>
                         <li>Anthropic</li>
                         <li>Google AI</li>
+                        <li>Google Vertex AI</li>
                         <li>Azure OpenAI</li>
                         <li>Ollama</li>
                         <li>OpenRouter</li>
                         <li>DeepSeek</li>
                         <li>SiliconFlow</li>
+                        <li>ModelScope</li>
                     </ul>
                     <p className="text-gray-700 mt-4">
                         注意：<code>claude-sonnet-4-5</code>{" "}

app/[lang]/about/ja/page.tsx

@@ -17,18 +17,7 @@ export const metadata: Metadata = {
     ],
 }
 
-function formatNumber(num: number): string {
-    if (num >= 1000) {
-        return `${num / 1000}k`
-    }
-    return num.toString()
-}
-
 export default function AboutJA() {
-    const dailyRequestLimit = Number(process.env.DAILY_REQUEST_LIMIT) || 20
-    const dailyTokenLimit = Number(process.env.DAILY_TOKEN_LIMIT) || 500000
-    const tpmLimit = Number(process.env.TPM_LIMIT) || 50000
-
     return (
         <div className="min-h-screen bg-gray-50">
             {/* Navigation */}
@@ -116,42 +105,6 @@ export default function AboutJA() {
                                 </p>
                             </div>
 
-                            {/* Usage Limits */}
-                            <p className="text-sm text-gray-600 mb-3">
-                                現在の使用制限：
-                            </p>
-                            <div className="grid grid-cols-3 gap-3 mb-5">
-                                <div className="text-center p-3 bg-white/60 rounded-lg">
-                                    <p className="text-lg font-bold text-amber-600">
-                                        {formatNumber(dailyRequestLimit)}
-                                    </p>
-                                    <p className="text-xs text-gray-500">
-                                        リクエスト/日
-                                    </p>
-                                </div>
-                                <div className="text-center p-3 bg-white/60 rounded-lg">
-                                    <p className="text-lg font-bold text-amber-600">
-                                        {formatNumber(dailyTokenLimit)}
-                                    </p>
-                                    <p className="text-xs text-gray-500">
-                                        トークン/日
-                                    </p>
-                                </div>
-                                <div className="text-center p-3 bg-white/60 rounded-lg">
-                                    <p className="text-lg font-bold text-amber-600">
-                                        {formatNumber(tpmLimit)}
-                                    </p>
-                                    <p className="text-xs text-gray-500">
-                                        トークン/分
-                                    </p>
-                                </div>
-                            </div>
-
-                            {/* Divider */}
-                            <div className="flex items-center gap-3 my-5">
-                                <div className="flex-1 h-px bg-gradient-to-r from-transparent via-amber-300 to-transparent" />
-                            </div>
-
                             {/* Bring Your Own Key */}
                             <div className="text-center">
                                 <h4 className="text-base font-bold text-gray-900 mb-2">
@@ -354,11 +307,13 @@ export default function AboutJA() {
                         </li>
                         <li>Anthropic</li>
                         <li>Google AI</li>
+                        <li>Google Vertex AI</li>
                         <li>Azure OpenAI</li>
                         <li>Ollama</li>
                         <li>OpenRouter</li>
                         <li>DeepSeek</li>
                         <li>SiliconFlow</li>
+                        <li>ModelScope</li>
                     </ul>
                     <p className="text-gray-700 mt-4">
                         注：<code>claude-sonnet-4-5</code>

app/[lang]/about/page.tsx

@@ -17,18 +17,7 @@ export const metadata: Metadata = {
     ],
 }
 
-function formatNumber(num: number): string {
-    if (num >= 1000) {
-        return `${num / 1000}k`
-    }
-    return num.toString()
-}
-
 export default function About() {
-    const dailyRequestLimit = Number(process.env.DAILY_REQUEST_LIMIT) || 20
-    const dailyTokenLimit = Number(process.env.DAILY_TOKEN_LIMIT) || 500000
-    const tpmLimit = Number(process.env.TPM_LIMIT) || 50000
-
     return (
         <div className="min-h-screen bg-gray-50">
             {/* Navigation */}
@@ -118,42 +107,6 @@ export default function About() {
                                 </p>
                             </div>
 
-                            {/* Usage Limits */}
-                            <p className="text-sm text-gray-600 mb-3">
-                                Please note the current usage limits:
-                            </p>
-                            <div className="grid grid-cols-3 gap-3 mb-5">
-                                <div className="text-center p-3 bg-white/60 rounded-lg">
-                                    <p className="text-lg font-bold text-amber-600">
-                                        {formatNumber(dailyRequestLimit)}
-                                    </p>
-                                    <p className="text-xs text-gray-500">
-                                        requests/day
-                                    </p>
-                                </div>
-                                <div className="text-center p-3 bg-white/60 rounded-lg">
-                                    <p className="text-lg font-bold text-amber-600">
-                                        {formatNumber(dailyTokenLimit)}
-                                    </p>
-                                    <p className="text-xs text-gray-500">
-                                        tokens/day
-                                    </p>
-                                </div>
-                                <div className="text-center p-3 bg-white/60 rounded-lg">
-                                    <p className="text-lg font-bold text-amber-600">
-                                        {formatNumber(tpmLimit)}
-                                    </p>
-                                    <p className="text-xs text-gray-500">
-                                        tokens/min
-                                    </p>
-                                </div>
-                            </div>
-
-                            {/* Divider */}
-                            <div className="flex items-center gap-3 my-5">
-                                <div className="flex-1 h-px bg-gradient-to-r from-transparent via-amber-300 to-transparent" />
-                            </div>
-
                             {/* Bring Your Own Key */}
                             <div className="text-center">
                                 <h4 className="text-base font-bold text-gray-900 mb-2">
@@ -373,11 +326,13 @@ export default function About() {
                         </li>
                         <li>Anthropic</li>
                         <li>Google AI</li>
+                        <li>Google Vertex AI</li>
                         <li>Azure OpenAI</li>
                         <li>Ollama</li>
                         <li>OpenRouter</li>
                         <li>DeepSeek</li>
                         <li>SiliconFlow</li>
+                        <li>ModelScope</li>
                     </ul>
                     <p className="text-gray-700 mt-4">
                         Note that <code>claude-sonnet-4-5</code> has trained on

app/[lang]/page.tsx

@@ -4,7 +4,6 @@ import { Suspense, useCallback, useEffect, useRef, useState } from "react"
 import { DrawIoEmbed } from "react-drawio"
 import type { ImperativePanelHandle } from "react-resizable-panels"
 import ChatPanel from "@/components/chat-panel"
-import { STORAGE_CLOSE_PROTECTION_KEY } from "@/components/settings-dialog"
 import {
     ResizableHandle,
     ResizablePanel,
@@ -29,7 +28,6 @@ export default function Home() {
     const [darkMode, setDarkMode] = useState(false)
     const [isLoaded, setIsLoaded] = useState(false)
     const [isDrawioReady, setIsDrawioReady] = useState(false)
-    const [closeProtection, setCloseProtection] = useState(false)
 
     const chatPanelRef = useRef<ImperativePanelHandle>(null)
     const isMobileRef = useRef(false)
@@ -66,13 +64,6 @@ export default function Home() {
             document.documentElement.classList.toggle("dark", prefersDark)
         }
 
-        const savedCloseProtection = localStorage.getItem(
-            STORAGE_CLOSE_PROTECTION_KEY,
-        )
-        if (savedCloseProtection === "true") {
-            setCloseProtection(true)
-        }
-
         setIsLoaded(true)
     }, [pathname, router])
 
@@ -146,20 +137,6 @@ export default function Home() {
         return () => window.removeEventListener("keydown", handleKeyDown)
     }, [])
 
-    // Show confirmation dialog when user tries to leave the page
-    useEffect(() => {
-        if (!closeProtection) return
-
-        const handleBeforeUnload = (event: BeforeUnloadEvent) => {
-            event.preventDefault()
-            return ""
-        }
-
-        window.addEventListener("beforeunload", handleBeforeUnload)
-        return () =>
-            window.removeEventListener("beforeunload", handleBeforeUnload)
-    }, [closeProtection])
-
     return (
         <div className="h-screen bg-background relative overflow-hidden">
             <ResizablePanelGroup
@@ -243,7 +220,6 @@ export default function Home() {
                                 darkMode={darkMode}
                                 onToggleDarkMode={handleDarkModeChange}
                                 isMobile={isMobile}
-                                onCloseProtectionChange={setCloseProtection}
                             />
                         </Suspense>
                     </div>

app/api/chat/route.ts

@@ -34,6 +34,7 @@ import {
     setTraceOutput,
     wrapWithObserve,
 } from "@/lib/langfuse"
+import { findServerModelById } from "@/lib/server-model-config"
 import { getSystemPrompt } from "@/lib/system-prompts"
 import { getUserIdFromRequest } from "@/lib/user-id"
 
@@ -168,6 +169,7 @@ async function handleChatRequest(req: Request): Promise<Response> {
     // Read client AI provider overrides from headers
     const provider = req.headers.get("x-ai-provider")
     let baseUrl = req.headers.get("x-ai-base-url")
+    const selectedModelId = req.headers.get("x-selected-model-id")
 
     // For EdgeOne provider, construct full URL from request origin
     // because createOpenAI needs absolute URL, not relative path
@@ -179,8 +181,30 @@ async function handleChatRequest(req: Request): Promise<Response> {
     // Get cookie header for EdgeOne authentication (eo_token, eo_time)
     const cookieHeader = req.headers.get("cookie")
 
+    // Check if this is a server model with custom env var names
+    let serverModelConfig: {
+        apiKeyEnv?: string
+        baseUrlEnv?: string
+        provider?: string
+    } = {}
+    if (selectedModelId?.startsWith("server:")) {
+        const serverModel = await findServerModelById(selectedModelId)
+        console.log(
+            `[Server Model Lookup] ID: ${selectedModelId}, Found: ${!!serverModel}, Provider: ${serverModel?.provider}`,
+        )
+        if (serverModel) {
+            serverModelConfig = {
+                apiKeyEnv: serverModel.apiKeyEnv,
+                baseUrlEnv: serverModel.baseUrlEnv,
+                // Use actual provider from config (client header may have incorrect value due to ID format change)
+                provider: serverModel.provider,
+            }
+        }
+    }
+
     const clientOverrides = {
-        provider,
+        // Server model provider takes precedence over client header
+        provider: serverModelConfig.provider || provider,
         baseUrl,
         apiKey: req.headers.get("x-ai-api-key"),
         modelId: req.headers.get("x-ai-model"),
@@ -189,6 +213,10 @@ async function handleChatRequest(req: Request): Promise<Response> {
         awsSecretAccessKey: req.headers.get("x-aws-secret-access-key"),
         awsRegion: req.headers.get("x-aws-region"),
         awsSessionToken: req.headers.get("x-aws-session-token"),
+        // Server model custom env var names
+        ...serverModelConfig,
+        // Vertex AI credentials (Express Mode)
+        vertexApiKey: req.headers.get("x-vertex-api-key"),
         // Pass cookies for EdgeOne Pages authentication
         ...(provider === "edgeone" &&
             cookieHeader && {
@@ -199,6 +227,10 @@ async function handleChatRequest(req: Request): Promise<Response> {
     // Read minimal style preference from header
     const minimalStyle = req.headers.get("x-minimal-style") === "true"
 
+    console.log(
+        `[Client Overrides] provider: ${clientOverrides.provider}, modelId: ${clientOverrides.modelId}`,
+    )
+
     // Get AI model with optional client overrides
     const { model, providerOptions, headers, modelId } =
         getAIModel(clientOverrides)
@@ -441,6 +473,13 @@ ${userInputText}
                         inputToRepair = inputToRepair.replace(/:=/g, ": ")
                         // Fix `= "` instead of `: "`
                         inputToRepair = inputToRepair.replace(/=\s*"/g, ': "')
+                        // Fix inconsistent quote escaping in XML attributes within JSON strings
+                        // Pattern: attribute="value\" where opening quote is unescaped but closing is escaped
+                        // Example: y="-20\" should be y=\"-20\"
+                        inputToRepair = inputToRepair.replace(
+                            /(\w+)="([^"]*?)\\"/g,
+                            '$1=\\"$2\\"',
+                        )
                     }
                     // Use jsonrepair to fix truncated JSON
                     const repairedInput = jsonrepair(inputToRepair)

app/api/parse-url/route.ts

@@ -0,0 +1,154 @@
+import { extract } from "@extractus/article-extractor"
+import { NextResponse } from "next/server"
+import TurndownService from "turndown"
+
+const MAX_CONTENT_LENGTH = 150000 // Match PDF limit
+const EXTRACT_TIMEOUT_MS = 15000
+
+// SSRF protection - block private/internal addresses
+function isPrivateUrl(urlString: string): boolean {
+    try {
+        const url = new URL(urlString)
+        const hostname = url.hostname.toLowerCase()
+
+        // Block localhost
+        if (
+            hostname === "localhost" ||
+            hostname === "127.0.0.1" ||
+            hostname === "::1"
+        ) {
+            return true
+        }
+
+        // Block AWS/cloud metadata endpoints
+        if (
+            hostname === "169.254.169.254" ||
+            hostname === "metadata.google.internal"
+        ) {
+            return true
+        }
+
+        // Check for private IPv4 ranges
+        const ipv4Match = hostname.match(
+            /^(\d{1,3})\.(\d{1,3})\.(\d{1,3})\.(\d{1,3})$/,
+        )
+        if (ipv4Match) {
+            const [, a, b] = ipv4Match.map(Number)
+            if (a === 10) return true // 10.0.0.0/8
+            if (a === 172 && b >= 16 && b <= 31) return true // 172.16.0.0/12
+            if (a === 192 && b === 168) return true // 192.168.0.0/16
+            if (a === 169 && b === 254) return true // 169.254.0.0/16 (link-local)
+            if (a === 127) return true // 127.0.0.0/8 (loopback)
+        }
+
+        // Block common internal hostnames
+        if (
+            hostname.endsWith(".local") ||
+            hostname.endsWith(".internal") ||
+            hostname.endsWith(".localhost")
+        ) {
+            return true
+        }
+
+        return false
+    } catch {
+        return true // Invalid URL - block it
+    }
+}
+
+export async function POST(req: Request) {
+    try {
+        const { url } = await req.json()
+
+        if (!url || typeof url !== "string") {
+            return NextResponse.json(
+                { error: "URL is required" },
+                { status: 400 },
+            )
+        }
+
+        // Validate URL format
+        try {
+            new URL(url)
+        } catch {
+            return NextResponse.json(
+                { error: "Invalid URL format" },
+                { status: 400 },
+            )
+        }
+
+        // SSRF protection
+        if (isPrivateUrl(url)) {
+            return NextResponse.json(
+                { error: "Cannot access private/internal URLs" },
+                { status: 400 },
+            )
+        }
+
+        // Extract article content with timeout to avoid tying up server resources
+        const controller = new AbortController()
+        const timeoutId = setTimeout(() => {
+            controller.abort()
+        }, EXTRACT_TIMEOUT_MS)
+
+        let article
+        try {
+            article = await extract(url, undefined, {
+                headers: {
+                    "User-Agent": "Mozilla/5.0 (compatible; NextAIDrawio/1.0)",
+                },
+                signal: controller.signal,
+            })
+        } catch (err: any) {
+            if (err?.name === "AbortError") {
+                return NextResponse.json(
+                    { error: "Timed out while fetching URL content" },
+                    { status: 504 },
+                )
+            }
+            throw err
+        } finally {
+            clearTimeout(timeoutId)
+        }
+
+        if (!article || !article.content) {
+            return NextResponse.json(
+                { error: "Could not extract content from URL" },
+                { status: 400 },
+            )
+        }
+
+        // Convert HTML to Markdown
+        const turndownService = new TurndownService({
+            headingStyle: "atx",
+            codeBlockStyle: "fenced",
+        })
+
+        // Remove unwanted elements before conversion
+        turndownService.remove(["script", "style", "iframe", "noscript"])
+
+        const markdown = turndownService.turndown(article.content)
+
+        // Check content length
+        if (markdown.length > MAX_CONTENT_LENGTH) {
+            return NextResponse.json(
+                {
+                    error: `Content exceeds ${MAX_CONTENT_LENGTH / 1000}k character limit (${(markdown.length / 1000).toFixed(1)}k chars)`,
+                },
+                { status: 400 },
+            )
+        }
+
+        return NextResponse.json({
+            title: article.title || "Untitled",
+            content: markdown,
+            charCount: markdown.length,
+        })
+    } catch (error) {
+        console.error("URL extraction error:", error)
+        return NextResponse.json(
+            { error: "Failed to fetch or parse URL content" },
+            { status: 500 },
+        )
+    }
+}

app/api/server-models/route.ts

@@ -0,0 +1,14 @@
+import { NextResponse } from "next/server"
+import { loadFlattenedServerModels } from "@/lib/server-model-config"
+
+// Use dynamic rendering to read AI_MODEL/AI_PROVIDER env vars at runtime
+// This ensures Docker users can set these values when starting containers
+export const dynamic = "force-dynamic"
+
+export async function GET() {
+    const models = await loadFlattenedServerModels()
+    return NextResponse.json({
+        models,
+        hasConfig: models.length > 0,
+    })
+}

app/api/validate-model/route.ts

@@ -3,6 +3,7 @@ import { createAnthropic } from "@ai-sdk/anthropic"
 import { createDeepSeek, deepseek } from "@ai-sdk/deepseek"
 import { createGateway } from "@ai-sdk/gateway"
 import { createGoogleGenerativeAI } from "@ai-sdk/google"
+import { createVertex } from "@ai-sdk/google-vertex"
 import { createOpenAI } from "@ai-sdk/openai"
 import { createOpenRouter } from "@openrouter/ai-sdk-provider"
 import { generateText } from "ai"
@@ -80,6 +81,8 @@ interface ValidateRequest {
     awsAccessKeyId?: string
     awsSecretAccessKey?: string
     awsRegion?: string
+    // Vertex AI specific
+    vertexApiKey?: string // Express Mode API key
 }
 
 export async function POST(req: Request) {
@@ -93,6 +96,8 @@ export async function POST(req: Request) {
             awsAccessKeyId,
             awsSecretAccessKey,
             awsRegion,
+            // Note: Express Mode only needs vertexApiKey
+            vertexApiKey,
         } = body
 
         if (!provider || !modelId) {
@@ -121,6 +126,16 @@ export async function POST(req: Request) {
                     { status: 400 },
                 )
             }
+        } else if (provider === "vertexai") {
+            if (!vertexApiKey) {
+                return NextResponse.json(
+                    {
+                        valid: false,
+                        error: "Vertex AI API key is required for Express Mode",
+                    },
+                    { status: 400 },
+                )
+            }
         } else if (provider !== "ollama" && provider !== "edgeone" && !apiKey) {
             return NextResponse.json(
                 { valid: false, error: "API key is required" },
@@ -158,6 +173,15 @@ export async function POST(req: Request) {
                 break
             }
 
+            case "vertexai": {
+                const vertex = createVertex({
+                    apiKey: vertexApiKey,
+                    ...(baseUrl && { baseURL: baseUrl }),
+                })
+                model = vertex(modelId)
+                break
+            }
+
             case "azure": {
                 const azure = createOpenAI({
                     apiKey,
@@ -202,7 +226,7 @@ export async function POST(req: Request) {
             case "siliconflow": {
                 const sf = createOpenAI({
                     apiKey,
-                    baseURL: baseUrl || "https://api.siliconflow.com/v1",
+                    baseURL: baseUrl || "https://api.siliconflow.cn/v1",
                 })
                 model = sf.chat(modelId)
                 break
@@ -251,16 +275,98 @@ export async function POST(req: Request) {
             }
 
             case "doubao": {
-                // ByteDance Doubao uses DeepSeek-compatible API
-                const doubao = createDeepSeek({
-                    apiKey,
-                    baseURL:
-                        baseUrl || "https://ark.cn-beijing.volces.com/api/v3",
-                })
-                model = doubao(modelId)
+                // ByteDance Doubao: use DeepSeek for DeepSeek/Kimi models, OpenAI for others
+                const doubaoBaseUrl =
+                    baseUrl || "https://ark.cn-beijing.volces.com/api/v3"
+                const lowerModelId = modelId.toLowerCase()
+                if (
+                    lowerModelId.includes("deepseek") ||
+                    lowerModelId.includes("kimi")
+                ) {
+                    const doubao = createDeepSeek({
+                        apiKey,
+                        baseURL: doubaoBaseUrl,
+                    })
+                    model = doubao(modelId)
+                } else {
+                    const doubao = createOpenAI({
+                        apiKey,
+                        baseURL: doubaoBaseUrl,
+                    })
+                    model = doubao.chat(modelId)
+                }
                 break
             }
 
+            case "modelscope": {
+                const baseURL =
+                    baseUrl || "https://api-inference.modelscope.cn/v1"
+                const startTime = Date.now()
+
+                try {
+                    // Initiate a streaming request (required for QwQ-32B and certain Qwen3 models)
+                    const response = await fetch(
+                        `${baseURL}/chat/completions`,
+                        {
+                            method: "POST",
+                            headers: {
+                                "Content-Type": "application/json",
+                                Authorization: `Bearer ${apiKey}`,
+                            },
+                            body: JSON.stringify({
+                                model: modelId,
+                                messages: [
+                                    { role: "user", content: "Say 'OK'" },
+                                ],
+                                max_tokens: 20,
+                                stream: true,
+                                enable_thinking: false,
+                            }),
+                        },
+                    )
+
+                    if (!response.ok) {
+                        const errorText = await response.text()
+                        throw new Error(
+                            `ModelScope API error (${response.status}): ${errorText}`,
+                        )
+                    }
+
+                    const contentType =
+                        response.headers.get("content-type") || ""
+                    const isValidStreamingResponse =
+                        response.status === 200 &&
+                        (contentType.includes("text/event-stream") ||
+                            contentType.includes("application/json"))
+
+                    if (!isValidStreamingResponse) {
+                        throw new Error(
+                            `Unexpected response format: ${contentType}`,
+                        )
+                    }
+
+                    const responseTime = Date.now() - startTime
+
+                    if (response.body) {
+                        response.body.cancel().catch(() => {
+                            /* Ignore cancellation errors */
+                        })
+                    }
+
+                    return NextResponse.json({
+                        valid: true,
+                        responseTime,
+                        note: "ModelScope model validated (using streaming API)",
+                    })
+                } catch (error) {
+                    console.error(
+                        "[validate-model] ModelScope validation failed:",
+                        error,
+                    )
+                    throw error
+                }
+            }
+
             default:
                 return NextResponse.json(
                     { valid: false, error: `Unknown provider: ${provider}` },

app/globals.css

@@ -244,6 +244,19 @@
     .scrollbar-thin::-webkit-scrollbar-thumb:hover {
         background-color: oklch(0.75 0.01 260);
     }
+
+    /* Dark mode scrollbar */
+    .dark .scrollbar-thin {
+        scrollbar-color: oklch(0.35 0.015 260) transparent;
+    }
+
+    .dark .scrollbar-thin::-webkit-scrollbar-thumb {
+        background-color: oklch(0.35 0.015 260);
+    }
+
+    .dark .scrollbar-thin::-webkit-scrollbar-thumb:hover {
+        background-color: oklch(0.45 0.015 260);
+    }
 }
 
 /* Smooth page transitions */

components/ai-elements/model-selector.tsx

@@ -169,3 +169,27 @@ export const ModelSelectorName = ({
 }: ModelSelectorNameProps) => (
     <span className={cn("flex-1 truncate text-left", className)} {...props} />
 )
+
+export type ModelSelectorSectionHeaderProps = {
+    icon: ReactNode
+    label: string
+    className?: string
+}
+
+export const ModelSelectorSectionHeader = ({
+    icon,
+    label,
+    className,
+}: ModelSelectorSectionHeaderProps) => (
+    <div
+        className={cn(
+            "flex items-center gap-2 px-2 py-1.5 text-xs font-semibold text-muted-foreground bg-muted/40 rounded-sm mx-1 mt-1",
+            className,
+        )}
+    >
+        <span className="[&>svg]:size-3.5" aria-hidden="true">
+            {icon}
+        </span>
+        <span>{label}</span>
+    </div>
+)

components/chat-input.tsx

@@ -4,6 +4,7 @@ import {
     Download,
     History,
     Image as ImageIcon,
+    Link,
     Loader2,
     Send,
 } from "lucide-react"
@@ -18,11 +19,15 @@ import { SaveDialog } from "@/components/save-dialog"
 
 import { Button } from "@/components/ui/button"
 import { Textarea } from "@/components/ui/textarea"
+import { UrlInputDialog } from "@/components/url-input-dialog"
 import { useDiagram } from "@/contexts/diagram-context"
 import { useDictionary } from "@/hooks/use-dictionary"
 import { formatMessage } from "@/lib/i18n/utils"
 import { isPdfFile, isTextFile } from "@/lib/pdf-utils"
+import { STORAGE_KEYS } from "@/lib/storage"
 import type { FlattenedModel } from "@/lib/types/model-config"
+import { extractUrlContent, type UrlData } from "@/lib/url-utils"
+import { isRealDiagram } from "@/lib/utils"
 import { FilePreviewList } from "./file-preview-list"
 
 const MAX_IMAGE_SIZE = 2 * 1024 * 1024 // 2MB
@@ -144,6 +149,8 @@ interface ChatInputProps {
         File,
         { text: string; charCount: number; isExtracting: boolean }
     >
+    urlData?: Map<string, UrlData>
+    onUrlChange?: (data: Map<string, UrlData>) => void
 
     sessionId?: string
     error?: Error | null
@@ -163,6 +170,8 @@ export function ChatInput({
     files = [],
     onFileChange = () => {},
     pdfData = new Map(),
+    urlData,
+    onUrlChange,
     sessionId,
     error = null,
     models = [],
@@ -173,6 +182,7 @@ export function ChatInput({
 }: ChatInputProps) {
     const dict = useDictionary()
     const {
+        chartXML,
         diagramHistory,
         saveDiagramToFile,
         showSaveDialog,
@@ -183,6 +193,9 @@ export function ChatInput({
     const fileInputRef = useRef<HTMLInputElement>(null)
     const [isDragging, setIsDragging] = useState(false)
     const [showHistory, setShowHistory] = useState(false)
+    const [showUrlDialog, setShowUrlDialog] = useState(false)
+    const [isExtractingUrl, setIsExtractingUrl] = useState(false)
+    const [sendShortcut, setSendShortcut] = useState("ctrl-enter")
     // Allow retry when there's an error (even if status is still "streaming" or "submitted")
     const isDisabled =
         (status === "streaming" || status === "submitted") && !error
@@ -199,13 +212,36 @@ export function ChatInput({
         adjustTextareaHeight()
     }, [input, adjustTextareaHeight])
 
+    // Load send shortcut preference from localStorage and listen for changes
+    useEffect(() => {
+        const stored = localStorage.getItem(STORAGE_KEYS.sendShortcut)
+        if (stored) setSendShortcut(stored)
+
+        const handleChange = (e: CustomEvent<string>) =>
+            setSendShortcut(e.detail)
+        window.addEventListener(
+            "sendShortcutChange",
+            handleChange as EventListener,
+        )
+        return () =>
+            window.removeEventListener(
+                "sendShortcutChange",
+                handleChange as EventListener,
+            )
+    }, [])
+
     const handleChange = (e: React.ChangeEvent<HTMLTextAreaElement>) => {
         onChange(e)
         adjustTextareaHeight()
     }
 
     const handleKeyDown = (e: React.KeyboardEvent) => {
-        if ((e.metaKey || e.ctrlKey) && e.key === "Enter") {
+        const shouldSend =
+            sendShortcut === "enter"
+                ? e.key === "Enter" && !e.shiftKey && !e.ctrlKey && !e.metaKey
+                : (e.metaKey || e.ctrlKey) && e.key === "Enter"
+
+        if (shouldSend) {
             e.preventDefault()
             const form = e.currentTarget.closest("form")
             if (form && input.trim() && !isDisabled) {
@@ -312,6 +348,50 @@ export function ChatInput({
         }
     }
 
+    const handleUrlExtract = async (url: string) => {
+        if (!onUrlChange) return
+
+        setIsExtractingUrl(true)
+
+        try {
+            const existing = urlData
+                ? new Map(urlData)
+                : new Map<string, UrlData>()
+            existing.set(url, {
+                url,
+                title: url,
+                content: "",
+                charCount: 0,
+                isExtracting: true,
+            })
+            onUrlChange(existing)
+
+            const data = await extractUrlContent(url)
+
+            const newUrlData = new Map(existing)
+            newUrlData.set(url, data)
+            onUrlChange(newUrlData)
+
+            setShowUrlDialog(false)
+        } catch (error) {
+            // Remove the URL from the data map on error
+            const newUrlData = urlData
+                ? new Map(urlData)
+                : new Map<string, UrlData>()
+            newUrlData.delete(url)
+            onUrlChange(newUrlData)
+            showErrorToast(
+                <span className="text-muted-foreground">
+                    {error instanceof Error
+                        ? error.message
+                        : "Failed to extract URL content"}
+                </span>,
+            )
+        } finally {
+            setIsExtractingUrl(false)
+        }
+    }
+
     return (
         <form
             onSubmit={onSubmit}
@@ -324,13 +404,23 @@ export function ChatInput({
             onDragLeave={handleDragLeave}
             onDrop={handleDrop}
         >
-            {/* File previews */}
-            {files.length > 0 && (
+            {/* File & URL previews */}
+            {(files.length > 0 || (urlData && urlData.size > 0)) && (
                 <div className="mb-3">
                     <FilePreviewList
                         files={files}
                         onRemoveFile={handleRemoveFile}
                         pdfData={pdfData}
+                        urlData={urlData}
+                        onRemoveUrl={
+                            onUrlChange
+                                ? (url) => {
+                                      const next = new Map(urlData)
+                                      next.delete(url)
+                                      onUrlChange(next)
+                                  }
+                                : undefined
+                        }
                     />
                 </div>
             )}
@@ -344,7 +434,7 @@ export function ChatInput({
                     placeholder={dict.chat.placeholder}
                     disabled={isDisabled}
                     aria-label="Chat input"
-                    className="min-h-[60px] max-h-[200px] resize-none border-0 bg-transparent px-4 py-3 text-sm focus-visible:ring-0 focus-visible:ring-offset-0 placeholder:text-muted-foreground/60"
+                    className="min-h-[60px] max-h-[200px] resize-none border-0 bg-transparent px-4 py-3 text-sm focus-visible:ring-0 focus-visible:ring-offset-0 placeholder:text-muted-foreground/60 scrollbar-thin"
                 />
 
                 <div className="flex items-center justify-end gap-1 px-3 py-2 border-t border-border/50">
@@ -366,7 +456,7 @@ export function ChatInput({
                             variant="ghost"
                             size="sm"
                             onClick={() => setShowSaveDialog(true)}
-                            disabled={isDisabled}
+                            disabled={isDisabled || !isRealDiagram(chartXML)}
                             tooltipContent={dict.chat.saveDiagram}
                             className="h-8 w-8 p-0 text-muted-foreground hover:text-foreground"
                         >
@@ -385,6 +475,20 @@ export function ChatInput({
                             <ImageIcon className="h-4 w-4" />
                         </ButtonWithTooltip>
 
+                        {onUrlChange && (
+                            <ButtonWithTooltip
+                                type="button"
+                                variant="ghost"
+                                size="sm"
+                                onClick={() => setShowUrlDialog(true)}
+                                disabled={isDisabled}
+                                tooltipContent={dict.chat.ExtractURL}
+                                className="h-8 w-8 p-0 text-muted-foreground hover:text-foreground"
+                            >
+                                <Link className="h-4 w-4" />
+                            </ButtonWithTooltip>
+                        )}
+
                         <input
                             type="file"
                             ref={fileInputRef}
@@ -443,6 +547,14 @@ export function ChatInput({
                     .toISOString()
                     .slice(0, 10)}`}
             />
+            {onUrlChange && (
+                <UrlInputDialog
+                    open={showUrlDialog}
+                    onOpenChange={setShowUrlDialog}
+                    onSubmit={handleUrlExtract}
+                    isExtracting={isExtractingUrl}
+                />
+            )}
         </form>
     )
 }

components/chat-message-display.tsx

@@ -1114,7 +1114,7 @@ export function ChatMessageDisplay({
                                                                                                         )}
                                                                                                     </button>
                                                                                                     {isExpanded && (
-                                                                                                        <div className="px-3 py-2 border-t border-border/40 max-h-48 overflow-y-auto bg-muted/30">
+                                                                                                        <div className="px-3 py-2 border-t border-border/40 max-h-48 overflow-y-auto bg-muted/30 scrollbar-thin">
                                                                                                             <pre className="text-xs whitespace-pre-wrap text-foreground/80">
                                                                                                                 {
                                                                                                                     section.content

components/chat-panel.tsx

@@ -34,6 +34,7 @@ import { findCachedResponse } from "@/lib/cached-responses"
 import { formatMessage } from "@/lib/i18n/utils"
 import { isPdfFile, isTextFile } from "@/lib/pdf-utils"
 import { sanitizeMessages } from "@/lib/session-storage"
+import type { UrlData } from "@/lib/url-utils"
 import { type FileData, useFileProcessor } from "@/lib/use-file-processor"
 import { useQuotaManager } from "@/lib/use-quota-manager"
 import { cn, formatXML, isRealDiagram } from "@/lib/utils"
@@ -69,7 +70,6 @@ interface ChatPanelProps {
     darkMode: boolean
     onToggleDarkMode: () => void
     isMobile?: boolean
-    onCloseProtectionChange?: (enabled: boolean) => void
 }
 
 // Constants for tool states
@@ -110,7 +110,6 @@ export default function ChatPanel({
     darkMode,
     onToggleDarkMode,
     isMobile = false,
-    onCloseProtectionChange,
 }: ChatPanelProps) {
     const {
         loadDiagram: onDisplayChart,
@@ -158,6 +157,7 @@ export default function ChatPanel({
 
     // File processing using extracted hook
     const { files, pdfData, handleFileChange, setFiles } = useFileProcessor()
+    const [urlData, setUrlData] = useState<Map<string, UrlData>>(new Map())
 
     const [showSettingsDialog, setShowSettingsDialog] = useState(false)
     const [showModelConfigDialog, setShowModelConfigDialog] = useState(false)
@@ -710,6 +710,8 @@ export default function ChatPanel({
                         input,
                         files,
                         pdfData,
+                        undefined,
+                        urlData,
                     )
 
                     setMessages([
@@ -735,6 +737,7 @@ export default function ChatPanel({
                     setInput("")
                     sessionStorage.removeItem(SESSION_STORAGE_INPUT_KEY)
                     setFiles([])
+                    setUrlData(new Map())
                     return
                 }
             }
@@ -755,6 +758,7 @@ export default function ChatPanel({
                     files,
                     pdfData,
                     parts,
+                    urlData,
                 )
 
                 // Add the combined text as the first part
@@ -779,6 +783,7 @@ export default function ChatPanel({
                 setInput("")
                 sessionStorage.removeItem(SESSION_STORAGE_INPUT_KEY)
                 setFiles([])
+                setUrlData(new Map())
             } catch (error) {
                 console.error("Error fetching chart data:", error)
             }
@@ -854,6 +859,7 @@ export default function ChatPanel({
         clearDiagram()
         setDiagramHistory([])
         handleFileChange([]) // Use handleFileChange to also clear pdfData
+        setUrlData(new Map())
         const newSessionId = `session-${Date.now()}-${Math.random()
             .toString(36)
             .slice(2, 9)}`
@@ -957,6 +963,14 @@ export default function ChatPanel({
                         ...(config.awsSessionToken && {
                             "x-aws-session-token": config.awsSessionToken,
                         }),
+                        // Vertex AI credentials (Express Mode)
+                        ...(config.vertexApiKey && {
+                            "x-vertex-api-key": config.vertexApiKey,
+                        }),
+                    }),
+                    // Send selected model ID for server model lookup (apiKeyEnv/baseUrlEnv)
+                    ...(config.selectedModelId && {
+                        "x-selected-model-id": config.selectedModelId,
                     }),
                     ...(minimalStyle && {
                         "x-minimal-style": "true",
@@ -972,6 +986,7 @@ export default function ChatPanel({
         files: File[],
         pdfData: Map<File, FileData>,
         imageParts?: any[],
+        urlDataParam?: Map<string, UrlData>,
     ): Promise<string> => {
         let userText = baseText
 
@@ -1002,6 +1017,14 @@ export default function ChatPanel({
             }
         }
 
+        if (urlDataParam) {
+            for (const [url, data] of urlDataParam) {
+                if (data.content) {
+                    userText += `\n\n[URL: ${url}]\nTitle: ${data.title}\n\n${data.content}`
+                }
+            }
+        }
+
         return userText
     }
 
@@ -1264,6 +1287,8 @@ export default function ChatPanel({
                     files={files}
                     onFileChange={handleFileChange}
                     pdfData={pdfData}
+                    urlData={urlData}
+                    onUrlChange={setUrlData}
                     sessionId={sessionId}
                     error={error}
                     models={modelConfig.models}
@@ -1277,7 +1302,6 @@ export default function ChatPanel({
             <SettingsDialog
                 open={showSettingsDialog}
                 onOpenChange={setShowSettingsDialog}
-                onCloseProtectionChange={onCloseProtectionChange}
                 drawioUi={drawioUi}
                 onToggleDrawioUi={onToggleDrawioUi}
                 darkMode={darkMode}

components/file-preview-list.tsx

@@ -1,6 +1,6 @@
 "use client"
 
-import { FileCode, FileText, Loader2, X } from "lucide-react"
+import { FileCode, FileText, Link, Loader2, X } from "lucide-react"
 import Image from "next/image"
 import { useEffect, useRef, useState } from "react"
 import { useDictionary } from "@/hooks/use-dictionary"
@@ -20,12 +20,19 @@ interface FilePreviewListProps {
         File,
         { text: string; charCount: number; isExtracting: boolean }
     >
+    urlData?: Map<
+        string,
+        { url: string; title: string; charCount: number; isExtracting: boolean }
+    >
+    onRemoveUrl?: (url: string) => void
 }
 
 export function FilePreviewList({
     files,
     onRemoveFile,
     pdfData = new Map(),
+    urlData,
+    onRemoveUrl,
 }: FilePreviewListProps) {
     const dict = useDictionary()
     const [selectedImage, setSelectedImage] = useState<string | null>(null)
@@ -77,7 +84,7 @@ export function FilePreviewList({
         }
     }, [imageUrls, selectedImage])
 
-    if (files.length === 0) return null
+    if (files.length === 0 && (!urlData || urlData.size === 0)) return null
 
     return (
         <>
@@ -152,6 +159,59 @@ export function FilePreviewList({
                         </div>
                     )
                 })}
+                {/* URL previews */}
+                {urlData && urlData.size > 0 && (
+                    <div className="flex flex-wrap gap-2">
+                        {Array.from(urlData.entries()).map(
+                            ([url, data], index) => (
+                                <div
+                                    key={url + index}
+                                    className="relative group"
+                                >
+                                    <div className="w-20 h-20 border rounded-md overflow-hidden bg-muted">
+                                        <div className="flex flex-col items-center justify-center h-full p-1">
+                                            {data.isExtracting ? (
+                                                <>
+                                                    <Loader2 className="h-6 w-6 text-blue-500 mb-1 animate-spin" />
+                                                    <span className="text-[10px] text-muted-foreground">
+                                                        {dict.file.reading}
+                                                    </span>
+                                                </>
+                                            ) : (
+                                                <>
+                                                    <Link className="h-6 w-6 text-blue-500 mb-1" />
+                                                    <span className="text-xs text-center truncate w-full px-1">
+                                                        {data.title.length > 10
+                                                            ? `${data.title.slice(0, 7)}...`
+                                                            : data.title}
+                                                    </span>
+                                                    {data.charCount && (
+                                                        <span className="text-[10px] text-green-600 font-medium">
+                                                            {formatCharCount(
+                                                                data.charCount,
+                                                            )}{" "}
+                                                            {dict.file.chars}
+                                                        </span>
+                                                    )}
+                                                </>
+                                            )}
+                                        </div>
+                                    </div>
+                                    {onRemoveUrl && (
+                                        <button
+                                            type="button"
+                                            onClick={() => onRemoveUrl(url)}
+                                            className="absolute -top-2 -right-2 bg-destructive rounded-full p-1 opacity-0 group-hover:opacity-100 transition-opacity"
+                                            aria-label={dict.file.removeFile}
+                                        >
+                                            <X className="h-3 w-3" />
+                                        </button>
+                                    )}
+                                </div>
+                            ),
+                        )}
+                    </div>
+                )}
             </div>
             {/* Image Modal/Lightbox */}
             {selectedImage && (

components/history-dialog.tsx

@@ -43,7 +43,7 @@ export function HistoryDialog({
 
     return (
         <Dialog open={showHistory} onOpenChange={onToggleHistory}>
-            <DialogContent className="max-w-3xl max-h-[80vh] overflow-y-auto">
+            <DialogContent className="max-w-3xl max-h-[80vh] overflow-y-auto scrollbar-thin">
                 <DialogHeader>
                     <DialogTitle>{dict.history.title}</DialogTitle>
                     <DialogDescription>

components/model-config-dialog.tsx

@@ -78,7 +78,9 @@ const PROVIDER_LOGO_MAP: Record<string, string> = {
     sglang: "openai", // SGLang is OpenAI-compatible
     gateway: "vercel",
     edgeone: "tencent-cloud",
+    vertexai: "google",
     doubao: "bytedance",
+    modelscope: "modelscope",
 }
 
 // Provider logo component
@@ -236,6 +238,7 @@ export function ModelConfigDialog({
             "awsAccessKeyId",
             "awsSecretAccessKey",
             "awsRegion",
+            "vertexApiKey",
         ]
         if (credentialFields.includes(field)) {
             setValidationStatus("idle")
@@ -279,6 +282,7 @@ export function ModelConfigDialog({
         // Check credentials based on provider type
         const isBedrock = selectedProvider.provider === "bedrock"
         const isEdgeOne = selectedProvider.provider === "edgeone"
+        const isVertexAI = selectedProvider.provider === "vertexai"
         if (isBedrock) {
             if (
                 !selectedProvider.awsAccessKeyId ||
@@ -287,6 +291,11 @@ export function ModelConfigDialog({
             ) {
                 return
             }
+        } else if (isVertexAI) {
+            // Vertex AI requires vertexApiKey for Express Mode
+            if (!selectedProvider.vertexApiKey) {
+                return
+            }
         } else if (!isEdgeOne && !selectedProvider.apiKey) {
             return
         }
@@ -327,6 +336,8 @@ export function ModelConfigDialog({
                         awsAccessKeyId: selectedProvider.awsAccessKeyId,
                         awsSecretAccessKey: selectedProvider.awsSecretAccessKey,
                         awsRegion: selectedProvider.awsRegion,
+                        // Vertex AI credentials (Express Mode)
+                        vertexApiKey: selectedProvider.vertexApiKey,
                     }),
                 })
                 const data = await response.json()
@@ -866,7 +877,153 @@ export function ModelConfigDialog({
                                                     </div>
                                                 </>
                                             ) : selectedProvider.provider ===
-                                              "edgeone" ? (
+                                              "vertexai" ? (
+                                                <>
+                                                    {/* Vertex AI API Key */}
+                                                    <div className="space-y-2">
+                                                        <Label
+                                                            htmlFor="vertex-api-key"
+                                                            className="text-xs font-medium flex items-center gap-1.5"
+                                                        >
+                                                            <Key className="h-3.5 w-3.5 text-muted-foreground" />
+                                                            API Key
+                                                        </Label>
+                                                        <div className="flex gap-2">
+                                                            <div className="relative flex-1">
+                                                                <Input
+                                                                    id="vertex-api-key"
+                                                                    type={
+                                                                        showApiKey
+                                                                            ? "text"
+                                                                            : "password"
+                                                                    }
+                                                                    value={
+                                                                        selectedProvider.vertexApiKey ||
+                                                                        ""
+                                                                    }
+                                                                    onChange={(
+                                                                        e,
+                                                                    ) =>
+                                                                        handleProviderUpdate(
+                                                                            "vertexApiKey",
+                                                                            e
+                                                                                .target
+                                                                                .value,
+                                                                        )
+                                                                    }
+                                                                    placeholder="Enter your Vertex AI API key"
+                                                                    className="h-9 pr-10 font-mono text-xs"
+                                                                />
+                                                                <button
+                                                                    type="button"
+                                                                    onClick={() =>
+                                                                        setShowApiKey(
+                                                                            !showApiKey,
+                                                                        )
+                                                                    }
+                                                                    aria-label={
+                                                                        showApiKey
+                                                                            ? "Hide API key"
+                                                                            : "Show API key"
+                                                                    }
+                                                                    className="absolute right-3 top-1/2 -translate-y-1/2 text-muted-foreground hover:text-foreground transition-colors focus-visible:outline-none focus-visible:ring-2 focus-visible:ring-ring focus-visible:ring-offset-2 rounded"
+                                                                >
+                                                                    {showApiKey ? (
+                                                                        <EyeOff className="h-4 w-4" />
+                                                                    ) : (
+                                                                        <Eye className="h-4 w-4" />
+                                                                    )}
+                                                                </button>
+                                                            </div>
+                                                            <Button
+                                                                variant={
+                                                                    validationStatus ===
+                                                                    "success"
+                                                                        ? "outline"
+                                                                        : "default"
+                                                                }
+                                                                size="sm"
+                                                                onClick={
+                                                                    handleValidate
+                                                                }
+                                                                disabled={
+                                                                    !selectedProvider.vertexApiKey ||
+                                                                    validationStatus ===
+                                                                        "validating"
+                                                                }
+                                                                className={cn(
+                                                                    "h-9 px-4",
+                                                                    validationStatus ===
+                                                                        "success" &&
+                                                                        "text-success border-success/30 bg-success-muted hover:bg-success-muted",
+                                                                )}
+                                                            >
+                                                                {validationStatus ===
+                                                                "validating" ? (
+                                                                    <Loader2 className="h-4 w-4 animate-spin" />
+                                                                ) : validationStatus ===
+                                                                  "success" ? (
+                                                                    <>
+                                                                        <Check className="h-4 w-4 mr-1.5 animate-check-pop" />
+                                                                        {
+                                                                            dict
+                                                                                .modelConfig
+                                                                                .verified
+                                                                        }
+                                                                    </>
+                                                                ) : (
+                                                                    dict
+                                                                        .modelConfig
+                                                                        .test
+                                                                )}
+                                                            </Button>
+                                                        </div>
+                                                        {validationStatus ===
+                                                            "error" &&
+                                                            validationError && (
+                                                                <p className="text-xs text-destructive flex items-center gap-1">
+                                                                    <X className="h-3 w-3" />
+                                                                    {
+                                                                        validationError
+                                                                    }
+                                                                </p>
+                                                            )}
+                                                    </div>
+
+                                                    {/* Base URL (optional) */}
+                                                    <div className="space-y-2">
+                                                        <Label
+                                                            htmlFor="vertex-base-url"
+                                                            className="text-xs font-medium flex items-center gap-1.5"
+                                                        >
+                                                            <Link2 className="h-3.5 w-3.5 text-muted-foreground" />
+                                                            Base URL{" "}
+                                                            <span className="text-muted-foreground font-normal">
+                                                                (optional)
+                                                            </span>
+                                                        </Label>
+                                                        <Input
+                                                            id="vertex-base-url"
+                                                            value={
+                                                                selectedProvider.baseUrl ||
+                                                                ""
+                                                            }
+                                                            onChange={(e) =>
+                                                                handleProviderUpdate(
+                                                                    "baseUrl",
+                                                                    e.target
+                                                                        .value,
+                                                                )
+                                                            }
+                                                            placeholder="Custom endpoint URL"
+                                                            className="h-9 font-mono text-xs"
+                                                        />
+                                                    </div>
+                                                </>
+                                            ) : selectedProvider.provider ===
+                                                  "ollama" ||
+                                              selectedProvider.provider ===
+                                                  "edgeone" ? (
                                                 <div className="space-y-3">
                                                     <div className="flex items-center gap-2">
                                                         <Button

components/model-selector.tsx

@@ -5,8 +5,10 @@ import {
     Bot,
     Check,
     ChevronDown,
+    Monitor,
     Server,
     Settings2,
+    User,
 } from "lucide-react"
 import { useEffect, useMemo, useRef, useState } from "react"
 import {
@@ -19,6 +21,7 @@ import {
     ModelSelectorLogo,
     ModelSelectorName,
     ModelSelector as ModelSelectorRoot,
+    ModelSelectorSectionHeader,
     ModelSelectorSeparator,
     ModelSelectorTrigger,
 } from "@/components/ai-elements/model-selector"
@@ -49,7 +52,9 @@ const PROVIDER_LOGO_MAP: Record<string, string> = {
     sglang: "openai", // SGLang is OpenAI-compatible, use OpenAI logo
     gateway: "vercel",
     edgeone: "tencent-cloud",
+    vertexai: "google",
     doubao: "bytedance",
+    modelscope: "modelscope",
 }
 
 // Group models by providerLabel (handles duplicate providers)
@@ -61,7 +66,11 @@ function groupModelsByProvider(
         { provider: string; models: FlattenedModel[] }
     >()
     for (const model of models) {
-        const key = model.providerLabel
+        // For server models, strip "Server · " prefix for cleaner grouping
+        const key =
+            model.source === "server"
+                ? model.providerLabel.replace(/^Server · /, "")
+                : model.providerLabel
         const existing = groups.get(key)
         if (existing) {
             existing.models.push(model)
@@ -89,10 +98,26 @@ export function ModelSelector({
         }
         return models.filter((m) => m.validated === true)
     }, [models, showUnvalidatedModels])
-    const groupedModels = useMemo(
-        () => groupModelsByProvider(displayModels),
+
+    // Separate server and user models
+    const serverModels = useMemo(
+        () => displayModels.filter((m) => m.source === "server"),
         [displayModels],
     )
+    const userModels = useMemo(
+        () => displayModels.filter((m) => m.source !== "server"),
+        [displayModels],
+    )
+
+    // Group each category separately
+    const groupedServerModels = useMemo(
+        () => groupModelsByProvider(serverModels),
+        [serverModels],
+    )
+    const groupedUserModels = useMemo(
+        () => groupModelsByProvider(userModels),
+        [userModels],
+    )
 
     // Find selected model for display
     const selectedModel = useMemo(
@@ -159,7 +184,7 @@ export function ModelSelector({
                         size="sm"
                         disabled={disabled}
                         className={cn(
-                            "hover:bg-accent gap-1.5 h-8 px-2 transition-all duration-150 ease-in-out",
+                            "hover:bg-accent gap-1.5 h-8 px-2 transition-[padding,background-color] duration-150 ease-in-out",
                             !showLabel && "px-1.5 justify-center",
                         )}
                         // accessibility: expose label to screen readers
@@ -196,83 +221,169 @@ export function ModelSelector({
                                 : dict.modelConfig.noModelsFound}
                         </ModelSelectorEmpty>
 
-                        {/* Server Default Option */}
-                        <ModelSelectorGroup heading={dict.modelConfig.default}>
-                            <ModelSelectorItem
-                                value="__server_default__"
-                                onSelect={handleSelect}
-                                className={cn(
-                                    "cursor-pointer",
-                                    !selectedModelId && "bg-accent",
-                                )}
+                        {/* Server Default Option - only show when no server models are configured */}
+                        {serverModels.length === 0 && (
+                            <ModelSelectorGroup
+                                heading={dict.modelConfig.default}
                             >
-                                <Check
+                                <ModelSelectorItem
+                                    value="__server_default__"
+                                    onSelect={handleSelect}
                                     className={cn(
-                                        "mr-2 h-4 w-4",
-                                        !selectedModelId
-                                            ? "opacity-100"
-                                            : "opacity-0",
+                                        "cursor-pointer",
+                                        !selectedModelId && "bg-accent",
                                     )}
-                                />
-                                <Server className="mr-2 h-4 w-4 text-muted-foreground" />
-                                <ModelSelectorName>
-                                    {dict.modelConfig.serverDefault}
-                                </ModelSelectorName>
-                            </ModelSelectorItem>
-                        </ModelSelectorGroup>
-
-                        {/* Configured Models by Provider */}
-                        {Array.from(groupedModels.entries()).map(
-                            ([
-                                providerLabel,
-                                { provider, models: providerModels },
-                            ]) => (
-                                <ModelSelectorGroup
-                                    key={providerLabel}
-                                    heading={providerLabel}
                                 >
-                                    {providerModels.map((model) => (
-                                        <ModelSelectorItem
-                                            key={model.id}
-                                            value={model.modelId}
-                                            onSelect={() =>
-                                                handleSelect(model.id)
-                                            }
-                                            className="cursor-pointer"
+                                    <Check
+                                        className={cn(
+                                            "mr-2 h-4 w-4",
+                                            !selectedModelId
+                                                ? "opacity-100"
+                                                : "opacity-0",
+                                        )}
+                                    />
+                                    <Server className="mr-2 h-4 w-4 text-muted-foreground" />
+                                    <ModelSelectorName>
+                                        {dict.modelConfig.serverDefault}
+                                    </ModelSelectorName>
+                                </ModelSelectorItem>
+                            </ModelSelectorGroup>
+                        )}
+
+                        {/* Server Models Section */}
+                        {serverModels.length > 0 && (
+                            <>
+                                <ModelSelectorSectionHeader
+                                    icon={<Monitor />}
+                                    label={dict.modelConfig.serverModels}
+                                />
+                                {Array.from(groupedServerModels.entries()).map(
+                                    ([
+                                        providerLabel,
+                                        { provider, models: providerModels },
+                                    ]) => (
+                                        <ModelSelectorGroup
+                                            key={`server-${providerLabel}`}
+                                            heading={providerLabel}
+                                            className="[&>[cmdk-group-heading]]:pl-4"
                                         >
-                                            <Check
-                                                className={cn(
-                                                    "mr-2 h-4 w-4",
-                                                    selectedModelId === model.id
-                                                        ? "opacity-100"
-                                                        : "opacity-0",
-                                                )}
-                                            />
-                                            <ModelSelectorLogo
-                                                provider={
-                                                    PROVIDER_LOGO_MAP[
-                                                        provider
-                                                    ] || provider
-                                                }
-                                                className="mr-2"
-                                            />
-                                            <ModelSelectorName>
-                                                {model.modelId}
-                                            </ModelSelectorName>
-                                            {model.validated !== true && (
-                                                <span
-                                                    title={
-                                                        dict.modelConfig
-                                                            .unvalidatedModelWarning
+                                            {providerModels.map((model) => (
+                                                <ModelSelectorItem
+                                                    key={model.id}
+                                                    value={model.modelId}
+                                                    onSelect={() =>
+                                                        handleSelect(model.id)
                                                     }
+                                                    className="cursor-pointer"
                                                 >
-                                                    <AlertTriangle className="ml-auto h-3 w-3 text-warning" />
-                                                </span>
-                                            )}
-                                        </ModelSelectorItem>
-                                    ))}
-                                </ModelSelectorGroup>
-                            ),
+                                                    <Check
+                                                        className={cn(
+                                                            "mr-2 h-4 w-4",
+                                                            selectedModelId ===
+                                                                model.id
+                                                                ? "opacity-100"
+                                                                : "opacity-0",
+                                                        )}
+                                                    />
+                                                    <ModelSelectorLogo
+                                                        provider={
+                                                            PROVIDER_LOGO_MAP[
+                                                                provider
+                                                            ] || provider
+                                                        }
+                                                        className="mr-2"
+                                                    />
+                                                    <ModelSelectorName>
+                                                        {model.modelId}
+                                                    </ModelSelectorName>
+                                                    {model.isDefault && (
+                                                        <span
+                                                            title={
+                                                                dict.modelConfig
+                                                                    .serverDefaultModel
+                                                            }
+                                                            className="ml-auto text-xs text-muted-foreground"
+                                                        >
+                                                            {
+                                                                dict.modelConfig
+                                                                    .default
+                                                            }
+                                                        </span>
+                                                    )}
+                                                </ModelSelectorItem>
+                                            ))}
+                                        </ModelSelectorGroup>
+                                    ),
+                                )}
+                            </>
+                        )}
+
+                        {/* User Models Section */}
+                        {userModels.length > 0 && (
+                            <>
+                                {serverModels.length > 0 && (
+                                    <ModelSelectorSeparator />
+                                )}
+                                <ModelSelectorSectionHeader
+                                    icon={<User />}
+                                    label={dict.modelConfig.userModels}
+                                />
+                                {Array.from(groupedUserModels.entries()).map(
+                                    ([
+                                        providerLabel,
+                                        { provider, models: providerModels },
+                                    ]) => (
+                                        <ModelSelectorGroup
+                                            key={`user-${providerLabel}`}
+                                            heading={providerLabel}
+                                            className="[&>[cmdk-group-heading]]:pl-4"
+                                        >
+                                            {providerModels.map((model) => (
+                                                <ModelSelectorItem
+                                                    key={model.id}
+                                                    value={model.modelId}
+                                                    onSelect={() =>
+                                                        handleSelect(model.id)
+                                                    }
+                                                    className="cursor-pointer"
+                                                >
+                                                    <Check
+                                                        className={cn(
+                                                            "mr-2 h-4 w-4",
+                                                            selectedModelId ===
+                                                                model.id
+                                                                ? "opacity-100"
+                                                                : "opacity-0",
+                                                        )}
+                                                    />
+                                                    <ModelSelectorLogo
+                                                        provider={
+                                                            PROVIDER_LOGO_MAP[
+                                                                provider
+                                                            ] || provider
+                                                        }
+                                                        className="mr-2"
+                                                    />
+                                                    <ModelSelectorName>
+                                                        {model.modelId}
+                                                    </ModelSelectorName>
+                                                    {model.validated !==
+                                                        true && (
+                                                        <span
+                                                            title={
+                                                                dict.modelConfig
+                                                                    .unvalidatedModelWarning
+                                                            }
+                                                        >
+                                                            <AlertTriangle className="ml-auto h-3 w-3 text-warning" />
+                                                        </span>
+                                                    )}
+                                                </ModelSelectorItem>
+                                            ))}
+                                        </ModelSelectorGroup>
+                                    ),
+                                )}
+                            </>
                         )}
 
                         {/* Configure Option */}
@@ -281,7 +392,7 @@ export function ModelSelector({
                             <ModelSelectorItem
                                 value="__configure__"
                                 onSelect={handleSelect}
-                                className="cursor-pointer"
+                                className="cursor-pointer text-muted-foreground hover:text-foreground"
                             >
                                 <Settings2 className="mr-2 h-4 w-4" />
                                 <ModelSelectorName>

components/settings-dialog.tsx

@@ -3,6 +3,7 @@
 import { Github, Info, Moon, Sun, Tag } from "lucide-react"
 import { usePathname, useRouter, useSearchParams } from "next/navigation"
 import { Suspense, useEffect, useState } from "react"
+import { toast } from "sonner"
 import { Button } from "@/components/ui/button"
 import {
     Dialog,
@@ -24,6 +25,7 @@ import { Switch } from "@/components/ui/switch"
 import { useDictionary } from "@/hooks/use-dictionary"
 import { getApiEndpoint } from "@/lib/base-path"
 import { i18n, type Locale } from "@/lib/i18n/config"
+import { STORAGE_KEYS } from "@/lib/storage"
 
 // Reusable setting item component for consistent layout
 function SettingItem({
@@ -59,7 +61,6 @@ const LANGUAGE_LABELS: Record<Locale, string> = {
 interface SettingsDialogProps {
     open: boolean
     onOpenChange: (open: boolean) => void
-    onCloseProtectionChange?: (enabled: boolean) => void
     drawioUi: "min" | "sketch"
     onToggleDrawioUi: () => void
     darkMode: boolean
@@ -69,7 +70,6 @@ interface SettingsDialogProps {
 }
 
 export const STORAGE_ACCESS_CODE_KEY = "next-ai-draw-io-access-code"
-export const STORAGE_CLOSE_PROTECTION_KEY = "next-ai-draw-io-close-protection"
 const STORAGE_ACCESS_CODE_REQUIRED_KEY = "next-ai-draw-io-access-code-required"
 
 function getStoredAccessCodeRequired(): boolean | null {
@@ -82,7 +82,6 @@ function getStoredAccessCodeRequired(): boolean | null {
 function SettingsContent({
     open,
     onOpenChange,
-    onCloseProtectionChange,
     drawioUi,
     onToggleDrawioUi,
     darkMode,
@@ -95,13 +94,18 @@ function SettingsContent({
     const pathname = usePathname() || "/"
     const search = useSearchParams()
     const [accessCode, setAccessCode] = useState("")
-    const [closeProtection, setCloseProtection] = useState(true)
     const [isVerifying, setIsVerifying] = useState(false)
     const [error, setError] = useState("")
     const [accessCodeRequired, setAccessCodeRequired] = useState(
         () => getStoredAccessCodeRequired() ?? false,
     )
     const [currentLang, setCurrentLang] = useState("en")
+    const [sendShortcut, setSendShortcut] = useState("ctrl-enter")
+
+    // Proxy settings state (Electron only)
+    const [httpProxy, setHttpProxy] = useState("")
+    const [httpsProxy, setHttpsProxy] = useState("")
+    const [isApplyingProxy, setIsApplyingProxy] = useState(false)
 
     useEffect(() => {
         // Only fetch if not cached in localStorage
@@ -143,13 +147,20 @@ function SettingsContent({
                 localStorage.getItem(STORAGE_ACCESS_CODE_KEY) || ""
             setAccessCode(storedCode)
 
-            const storedCloseProtection = localStorage.getItem(
-                STORAGE_CLOSE_PROTECTION_KEY,
+            const storedSendShortcut = localStorage.getItem(
+                STORAGE_KEYS.sendShortcut,
             )
-            // Default to true if not set
-            setCloseProtection(storedCloseProtection !== "false")
+            setSendShortcut(storedSendShortcut || "ctrl-enter")
 
             setError("")
+
+            // Load proxy settings (Electron only)
+            if (window.electronAPI?.getProxy) {
+                window.electronAPI.getProxy().then((config) => {
+                    setHttpProxy(config.httpProxy || "")
+                    setHttpsProxy(config.httpsProxy || "")
+                })
+            }
         }
     }, [open])
 
@@ -208,6 +219,46 @@ function SettingsContent({
         }
     }
 
+    const handleApplyProxy = async () => {
+        if (!window.electronAPI?.setProxy) return
+
+        // Validate proxy URLs (must start with http:// or https://)
+        const validateProxyUrl = (url: string): boolean => {
+            if (!url) return true // Empty is OK
+            return url.startsWith("http://") || url.startsWith("https://")
+        }
+
+        const trimmedHttp = httpProxy.trim()
+        const trimmedHttps = httpsProxy.trim()
+
+        if (trimmedHttp && !validateProxyUrl(trimmedHttp)) {
+            toast.error("HTTP Proxy must start with http:// or https://")
+            return
+        }
+        if (trimmedHttps && !validateProxyUrl(trimmedHttps)) {
+            toast.error("HTTPS Proxy must start with http:// or https://")
+            return
+        }
+
+        setIsApplyingProxy(true)
+        try {
+            const result = await window.electronAPI.setProxy({
+                httpProxy: trimmedHttp || undefined,
+                httpsProxy: trimmedHttps || undefined,
+            })
+
+            if (result.success) {
+                toast.success(dict.settings.proxyApplied)
+            } else {
+                toast.error(result.error || "Failed to apply proxy settings")
+            }
+        } catch {
+            toast.error("Failed to apply proxy settings")
+        } finally {
+            setIsApplyingProxy(false)
+        }
+    }
+
     return (
         <DialogContent className="sm:max-w-lg p-0 gap-0">
             {/* Header */}
@@ -333,25 +384,6 @@ function SettingsContent({
                         </Button>
                     </SettingItem>
 
-                    {/* Close Protection */}
-                    <SettingItem
-                        label={dict.settings.closeProtection}
-                        description={dict.settings.closeProtectionDescription}
-                    >
-                        <Switch
-                            id="close-protection"
-                            checked={closeProtection}
-                            onCheckedChange={(checked) => {
-                                setCloseProtection(checked)
-                                localStorage.setItem(
-                                    STORAGE_CLOSE_PROTECTION_KEY,
-                                    checked.toString(),
-                                )
-                                onCloseProtectionChange?.(checked)
-                            }}
-                        />
-                    </SettingItem>
-
                     {/* Diagram Style */}
                     <SettingItem
                         label={dict.settings.diagramStyle}
@@ -370,6 +402,91 @@ function SettingsContent({
                             </span>
                         </div>
                     </SettingItem>
+
+                    {/* Send Shortcut */}
+                    <SettingItem
+                        label={dict.settings.sendShortcut}
+                        description={dict.settings.sendShortcutDescription}
+                    >
+                        <Select
+                            value={sendShortcut}
+                            onValueChange={(value) => {
+                                setSendShortcut(value)
+                                localStorage.setItem(
+                                    STORAGE_KEYS.sendShortcut,
+                                    value,
+                                )
+                                window.dispatchEvent(
+                                    new CustomEvent("sendShortcutChange", {
+                                        detail: value,
+                                    }),
+                                )
+                            }}
+                        >
+                            <SelectTrigger
+                                id="send-shortcut-select"
+                                className="w-[170px] h-9 rounded-xl"
+                            >
+                                <SelectValue />
+                            </SelectTrigger>
+                            <SelectContent>
+                                <SelectItem value="enter">
+                                    {dict.settings.enterToSend}
+                                </SelectItem>
+                                <SelectItem value="ctrl-enter">
+                                    {dict.settings.ctrlEnterToSend}
+                                </SelectItem>
+                            </SelectContent>
+                        </Select>
+                    </SettingItem>
+
+                    {/* Proxy Settings - Electron only */}
+                    {typeof window !== "undefined" &&
+                        window.electronAPI?.isElectron && (
+                            <div className="py-4 space-y-3">
+                                <div className="space-y-0.5">
+                                    <Label className="text-sm font-medium">
+                                        {dict.settings.proxy}
+                                    </Label>
+                                    <p className="text-xs text-muted-foreground">
+                                        {dict.settings.proxyDescription}
+                                    </p>
+                                </div>
+
+                                <div className="space-y-2">
+                                    <Input
+                                        id="http-proxy"
+                                        type="text"
+                                        value={httpProxy}
+                                        onChange={(e) =>
+                                            setHttpProxy(e.target.value)
+                                        }
+                                        placeholder={`${dict.settings.httpProxy}: http://proxy:8080`}
+                                        className="h-9"
+                                    />
+                                    <Input
+                                        id="https-proxy"
+                                        type="text"
+                                        value={httpsProxy}
+                                        onChange={(e) =>
+                                            setHttpsProxy(e.target.value)
+                                        }
+                                        placeholder={`${dict.settings.httpsProxy}: http://proxy:8080`}
+                                        className="h-9"
+                                    />
+                                </div>
+
+                                <Button
+                                    onClick={handleApplyProxy}
+                                    disabled={isApplyingProxy}
+                                    className="h-9 px-4 rounded-xl w-full"
+                                >
+                                    {isApplyingProxy
+                                        ? "..."
+                                        : dict.settings.applyProxy}
+                                </Button>
+                            </div>
+                        )}
                 </div>
             </div>
 

components/ui/card.tsx

@@ -1,92 +0,0 @@
-import * as React from "react"
-
-import { cn } from "@/lib/utils"
-
-function Card({ className, ...props }: React.ComponentProps<"div">) {
-  return (
-    <div
-      data-slot="card"
-      className={cn(
-        "bg-card text-card-foreground flex flex-col gap-6 rounded-xl border py-6 shadow-sm",
-        className
-      )}
-      {...props}
-    />
-  )
-}
-
-function CardHeader({ className, ...props }: React.ComponentProps<"div">) {
-  return (
-    <div
-      data-slot="card-header"
-      className={cn(
-        "@container/card-header grid auto-rows-min grid-rows-[auto_auto] items-start gap-1.5 px-6 has-[data-slot=card-action]:grid-cols-[1fr_auto] [.border-b]:pb-6",
-        className
-      )}
-      {...props}
-    />
-  )
-}
-
-function CardTitle({ className, ...props }: React.ComponentProps<"div">) {
-  return (
-    <div
-      data-slot="card-title"
-      className={cn("leading-none font-semibold", className)}
-      {...props}
-    />
-  )
-}
-
-function CardDescription({ className, ...props }: React.ComponentProps<"div">) {
-  return (
-    <div
-      data-slot="card-description"
-      className={cn("text-muted-foreground text-sm", className)}
-      {...props}
-    />
-  )
-}
-
-function CardAction({ className, ...props }: React.ComponentProps<"div">) {
-  return (
-    <div
-      data-slot="card-action"
-      className={cn(
-        "col-start-2 row-span-2 row-start-1 self-start justify-self-end",
-        className
-      )}
-      {...props}
-    />
-  )
-}
-
-function CardContent({ className, ...props }: React.ComponentProps<"div">) {
-  return (
-    <div
-      data-slot="card-content"
-      className={cn("px-6", className)}
-      {...props}
-    />
-  )
-}
-
-function CardFooter({ className, ...props }: React.ComponentProps<"div">) {
-  return (
-    <div
-      data-slot="card-footer"
-      className={cn("flex items-center px-6 [.border-t]:pt-6", className)}
-      {...props}
-    />
-  )
-}
-
-export {
-  Card,
-  CardHeader,
-  CardFooter,
-  CardTitle,
-  CardAction,
-  CardDescription,
-  CardContent,
-}

components/url-input-dialog.tsx

@@ -0,0 +1,116 @@
+"use client"
+
+import { Link, Loader2 } from "lucide-react"
+import { useState } from "react"
+import { Button } from "@/components/ui/button"
+import {
+    Dialog,
+    DialogContent,
+    DialogDescription,
+    DialogFooter,
+    DialogHeader,
+    DialogTitle,
+} from "@/components/ui/dialog"
+import { Input } from "@/components/ui/input"
+import { useDictionary } from "@/hooks/use-dictionary"
+
+interface UrlInputDialogProps {
+    open: boolean
+    onOpenChange: (open: boolean) => void
+    onSubmit: (url: string) => void
+    isExtracting: boolean
+}
+
+export function UrlInputDialog({
+    open,
+    onOpenChange,
+    onSubmit,
+    isExtracting,
+}: UrlInputDialogProps) {
+    const dict = useDictionary()
+    const [url, setUrl] = useState("")
+    const [error, setError] = useState("")
+
+    const handleSubmit = () => {
+        setError("")
+
+        if (!url.trim()) {
+            setError(dict.url.enterUrl)
+            return
+        }
+
+        try {
+            new URL(url)
+        } catch {
+            setError(dict.url.invalidFormat)
+            return
+        }
+
+        onSubmit(url.trim())
+    }
+
+    const handleKeyDown = (e: React.KeyboardEvent<HTMLInputElement>) => {
+        if (e.key === "Enter" && !isExtracting) {
+            e.preventDefault()
+            handleSubmit()
+        }
+    }
+
+    return (
+        <Dialog open={open} onOpenChange={onOpenChange}>
+            <DialogContent className="sm:max-w-md">
+                <DialogHeader>
+                    <DialogTitle>{dict.url.title}</DialogTitle>
+                    <DialogDescription>
+                        {dict.url.description}
+                    </DialogDescription>
+                </DialogHeader>
+
+                <div className="space-y-4">
+                    <div className="space-y-2">
+                        <Input
+                            value={url}
+                            onChange={(e) => {
+                                setUrl(e.target.value)
+                                setError("")
+                            }}
+                            onKeyDown={handleKeyDown}
+                            placeholder="https://example.com/article"
+                            disabled={isExtracting}
+                            autoFocus
+                        />
+                        {error && (
+                            <p className="text-sm text-destructive">{error}</p>
+                        )}
+                    </div>
+                </div>
+
+                <DialogFooter>
+                    <Button
+                        variant="outline"
+                        onClick={() => onOpenChange(false)}
+                        disabled={isExtracting}
+                    >
+                        {dict.url.Cancel}
+                    </Button>
+                    <Button
+                        onClick={handleSubmit}
+                        disabled={isExtracting || !url.trim()}
+                    >
+                        {isExtracting ? (
+                            <>
+                                <Loader2 className="mr-2 h-4 w-4 animate-spin" />
+                                {dict.url.Extracting}
+                            </>
+                        ) : (
+                            <>
+                                <Link className="mr-2 h-4 w-4" />
+                                {dict.url.extract}
+                            </>
+                        )}
+                    </Button>
+                </DialogFooter>
+            </DialogContent>
+        </Dialog>
+    )
+}

docs/cn/FAQ.md

@@ -0,0 +1,78 @@
+# 常见问题解答 (FAQ)
+
+---
+
+## 1. 无法导出 PDF
+
+**问题**: Web 版点击导出 PDF 后跳转到 `convert.diagrams.net/node/export` 然后无响应
+
+**原因**: 嵌入式 Draw.io 不支持直接 PDF 导出，依赖外部转换服务，在 iframe 中无法正常工作
+
+**解决方案**: 先导出为图片（PNG），再打印转成 PDF
+
+**相关 Issue**: #539, #125
+
+---
+
+## 2. 无法访问 embed.diagrams.net（离线/内网部署）
+
+**问题**: 内网环境提示"找不到 embed.diagrams.net 的服务器 IP 地址"
+
+**关键点**: `NEXT_PUBLIC_*` 环境变量是**构建时**变量，会被打包到 JS 代码中，**运行时设置无效**！
+
+**解决方案**: 必须在构建时通过 `args` 传入：
+
+```yaml
+# docker-compose.yml
+services:
+  drawio:
+    image: jgraph/drawio:latest
+    ports: ["8080:8080"]
+  next-ai-draw-io:
+    build:
+      context: .
+      args:
+        - NEXT_PUBLIC_DRAWIO_BASE_URL=http://你的服务器IP:8080/
+    ports: ["3000:3000"]
+    env_file: .env
+```
+
+**内网用户**: 在外网修改 Dockerfile 并构建镜像，再传到内网使用
+
+**相关 Issue**: #295, #317
+
+---
+
+## 3. 自建模型只思考不画图
+
+**问题**: 本地部署的模型（如 Qwen、LiteLLM）只输出思考过程，不生成图表
+
+**可能原因**:
+1. **模型太小** - 小模型难以正确遵循 tool calling 指令，建议使用 32B+ 参数的模型
+2. **未开启 tool calling** - 模型服务需要配置 tool use 功能
+
+**解决方案**: 开启 tool calling，例如 vLLM：
+```bash
+python -m vllm.entrypoints.openai.api_server \
+    --model Qwen/Qwen3-32B \
+    --enable-auto-tool-choice \
+    --tool-call-parser hermes
+```
+
+**相关 Issue**: #269, #75
+
+---
+
+## 4. 上传图片后提示"未提供图片"
+
+**问题**: 上传图片后，系统显示"未提供图片"错误
+
+**可能原因**:
+1. 模型不支持视觉功能（如 Kimi K2、DeepSeek、Qwen 文本模型）
+
+**解决方案**:
+- 使用支持视觉的模型：GPT-5.2、Claude 4.5 Sonnet、Gemini 3 Pro
+- 模型名带 `vision` 或 `vl` 的支持图片
+- 更新到最新版本（v0.4.9+）
+
+**相关 Issue**: #324, #421, #469

docs/cn/README_CN.md

@@ -37,11 +37,12 @@ https://github.com/user-attachments/assets/b2eef5f3-b335-4e71-a755-dc2e80931979
     - [安装](#安装)
   - [部署](#部署)
     - [部署到腾讯云EdgeOne Pages](#部署到腾讯云edgeone-pages)
-    - [部署到Vercel（推荐）](#部署到vercel推荐)
+    - [部署到Vercel](#部署到vercel)
     - [部署到Cloudflare Workers](#部署到cloudflare-workers)
   - [多提供商支持](#多提供商支持)
   - [工作原理](#工作原理)
   - [支持与联系](#支持与联系)
+  - [常见问题](#常见问题)
   - [Star历史](#star历史)
 
 ## 示例
@@ -179,7 +180,7 @@ npm run dev
 
 同时，通过腾讯云EdgeOne Pages部署，也会获得[每日免费的DeepSeek模型额度](https://edgeone.cloud.tencent.com/pages/document/169925463311781888)。
 
-### 部署到Vercel（推荐）
+### 部署到Vercel
 
 [![Deploy with Vercel](https://vercel.com/button)](https://vercel.com/new/clone?repository-url=https%3A%2F%2Fgithub.com%2FDayuanJiang%2Fnext-ai-draw-io)
 
@@ -199,11 +200,13 @@ npm run dev
 -   OpenAI
 -   Anthropic
 -   Google AI
+-   Google Vertex AI
 -   Azure OpenAI
 -   Ollama
 -   OpenRouter
 -   DeepSeek
 -   SiliconFlow
+-   ModelScope
 -   SGLang
 -   Vercel AI Gateway
 
@@ -211,6 +214,10 @@ npm run dev
 
 📖 **[详细的提供商配置指南](./ai-providers.md)** - 查看各提供商的设置说明。
 
+### 服务端多模型配置
+
+管理员可以配置多个服务端模型，让所有用户无需提供个人 API Key 即可使用。通过 `AI_MODELS_CONFIG` 环境变量（JSON 字符串）或 `ai-models.json` 文件配置。
+
 **模型要求**：此任务需要强大的模型能力，因为它涉及生成具有严格格式约束的长文本（draw.io XML）。推荐使用 Claude Sonnet 4.5、GPT-5.1、Gemini 3 Pro 和 DeepSeek V3.2/R1。
 
 注意：`claude` 系列已在带有 AWS、Azure、GCP 等云架构 Logo 的 draw.io 图表上进行训练，因此如果您想创建云架构图，这是最佳选择。
@@ -237,6 +244,10 @@ npm run dev
 
 -   邮箱：me[at]jiang.jp
 
+## 常见问题
+
+请参阅 [FAQ](./FAQ.md) 了解常见问题和解决方案。
+
 ## Star历史
 
 [![Star History Chart](https://api.star-history.com/svg?repos=DayuanJiang/next-ai-draw-io&type=date&legend=top-left)](https://www.star-history.com/#DayuanJiang/next-ai-draw-io&type=date&legend=top-left)

docs/cn/ai-providers.md

@@ -152,6 +152,19 @@ AI_PROVIDER=ollama
 AI_MODEL=llama3.2
 ```
 
+### ModelScope
+
+```bash
+MODELSCOPE_API_KEY=your_api_key
+AI_MODEL=Qwen/Qwen3-235B-A22B-Instruct-2507
+```
+
+可选的自定义端点：
+
+```bash
+MODELSCOPE_BASE_URL=https://your-custom-endpoint
+```
+
 可选的自定义 URL：
 
 ```bash
@@ -204,6 +217,63 @@ AI_MODEL=openai/gpt-4o
 AI_PROVIDER=google  # 或：openai, anthropic, deepseek, siliconflow, doubao, azure, bedrock, openrouter, ollama, gateway, sglang
 ```
 
+## 服务端多模型配置
+
+管理员可以配置多个服务端模型，让所有用户无需提供个人 API Key 即可使用。
+
+### 配置方式
+
+**方式一：环境变量**（推荐用于云部署）
+
+设置 `AI_MODELS_CONFIG` 为 JSON 字符串：
+
+```bash
+AI_MODELS_CONFIG='{"providers":[{"name":"OpenAI","provider":"openai","models":["gpt-4o"],"default":true}]}'
+```
+
+**方式二：配置文件**
+
+在项目根目录创建 `ai-models.json` 文件（或通过 `AI_MODELS_CONFIG_PATH` 指定路径）。
+
+### 配置示例
+
+```json
+{
+  "providers": [
+    {
+      "name": "OpenAI Production",
+      "provider": "openai",
+      "models": ["gpt-4o", "gpt-4o-mini"],
+      "default": true
+    },
+    {
+      "name": "Custom DeepSeek",
+      "provider": "deepseek",
+      "models": ["deepseek-chat"],
+      "apiKeyEnv": "MY_DEEPSEEK_KEY",
+      "baseUrlEnv": "MY_DEEPSEEK_URL"
+    }
+  ]
+}
+```
+
+### 字段说明
+
+| 字段 | 必填 | 说明 |
+|------|------|------|
+| `name` | 是 | 显示名称（支持同一提供商多个配置） |
+| `provider` | 是 | 提供商类型（`openai`, `anthropic`, `google`, `bedrock` 等） |
+| `models` | 是 | 模型 ID 列表 |
+| `default` | 否 | 设为 `true` 表示默认选中该提供商的第一个模型 |
+| `apiKeyEnv` | 否 | 自定义 API Key 环境变量名（默认使用提供商标准变量如 `OPENAI_API_KEY`） |
+| `baseUrlEnv` | 否 | 自定义 Base URL 环境变量名 |
+
+### 说明
+
+- API Key 和凭证通过环境变量提供。默认使用标准变量名（如 `OPENAI_API_KEY`），也可通过 `apiKeyEnv` 指定自定义变量名。
+- `name` 字段允许同一提供商多个配置（例如 "OpenAI Production" 和 "OpenAI Staging" 都使用 `provider: "openai"` 但 `apiKeyEnv` 不同）。
+- 如果配置不存在，应用会回退到 `AI_PROVIDER`/`AI_MODEL` 环境变量配置。
+
 ## 模型能力要求
 
 此任务对模型能力要求极高，因为它涉及生成具有严格格式约束（draw.io XML）的长文本。

docs/en/FAQ.md

@@ -0,0 +1,78 @@
+# Frequently Asked Questions (FAQ)
+
+---
+
+## 1. Cannot Export PDF
+
+**Problem**: Web version redirects to `convert.diagrams.net/node/export` when exporting PDF, then nothing happens
+
+**Cause**: Embedded Draw.io doesn't support direct PDF export, it relies on external conversion service which doesn't work in iframe
+
+**Solution**: Export as image (PNG) first, then print to PDF
+
+**Related Issues**: #539, #125
+
+---
+
+## 2. Cannot Access embed.diagrams.net (Offline/Intranet Deployment)
+
+**Problem**: Intranet environment shows "Cannot find server IP address for embed.diagrams.net"
+
+**Key Point**: `NEXT_PUBLIC_*` environment variables are **build-time** variables, they get bundled into JS code. **Runtime settings don't work!**
+
+**Solution**: Must pass via `args` at build time:
+
+```yaml
+# docker-compose.yml
+services:
+  drawio:
+    image: jgraph/drawio:latest
+    ports: ["8080:8080"]
+  next-ai-draw-io:
+    build:
+      context: .
+      args:
+        - NEXT_PUBLIC_DRAWIO_BASE_URL=http://your-server-ip:8080/
+    ports: ["3000:3000"]
+    env_file: .env
+```
+
+**Intranet Users**: Modify Dockerfile and build image on external network, then transfer to intranet
+
+**Related Issues**: #295, #317
+
+---
+
+## 3. Self-hosted Model Only Thinks But Doesn't Draw
+
+**Problem**: Locally deployed models (e.g., Qwen, LiteLLM) only output thinking process, don't generate diagrams
+
+**Possible Causes**:
+1. **Model too small** - Small models struggle to follow tool calling instructions correctly, recommend 32B+ parameter models
+2. **Tool calling not enabled** - Model service needs tool use configuration
+
+**Solution**: Enable tool calling, e.g., vLLM:
+```bash
+python -m vllm.entrypoints.openai.api_server \
+    --model Qwen/Qwen3-32B \
+    --enable-auto-tool-choice \
+    --tool-call-parser hermes
+```
+
+**Related Issues**: #269, #75
+
+---
+
+## 4. "No Image Provided" After Uploading Image
+
+**Problem**: After uploading an image, the system shows "No image provided" error
+
+**Possible Causes**:
+1. Model doesn't support vision (e.g., Kimi K2, DeepSeek, Qwen text models)
+
+**Solution**:
+- Use vision-capable models: GPT-5.2, Claude 4.5 Sonnet, Gemini 3 Pro
+- Models with `vision` or `vl` in name support images
+- Update to latest version (v0.4.9+)
+
+**Related Issues**: #324, #421, #469

docs/en/ai-providers.md

@@ -33,6 +33,21 @@ Optional custom endpoint:
 GOOGLE_BASE_URL=https://your-custom-endpoint
 ```
 
+### Google Vertex AI (Enterprise GCP)
+
+Google Vertex AI offers enterprise-grade features and data residency. **Express Mode** allows for simple API key authentication, making it compatible with edge runtimes like Vercel and Cloudflare.
+
+```bash
+GOOGLE_VERTEX_API_KEY=your_api_key
+AI_MODEL=gemini-2.0-flash
+```
+
+Optional custom endpoint:
+
+```bash
+GOOGLE_VERTEX_BASE_URL=https://your-custom-endpoint
+```
+
 ### OpenAI
 
 ```bash
@@ -158,6 +173,19 @@ Optional custom URL:
 OLLAMA_BASE_URL=http://localhost:11434
 ```
 
+### ModelScope
+
+```bash
+MODELSCOPE_API_KEY=your_api_key
+AI_MODEL=Qwen/Qwen3-235B-A22B-Instruct-2507
+```
+
+Optional custom endpoint:
+
+```bash
+MODELSCOPE_BASE_URL=https://your-custom-endpoint
+```
+
 ### Vercel AI Gateway
 
 Vercel AI Gateway provides unified access to multiple AI providers through a single API key. This simplifies authentication and allows you to switch between providers without managing multiple API keys.
@@ -201,9 +229,66 @@ If you only configure **one** provider's API key, the system will automatically
 If you configure **multiple** API keys, you must explicitly set `AI_PROVIDER`:
 
 ```bash
-AI_PROVIDER=google  # or: openai, anthropic, deepseek, siliconflow, doubao, azure, bedrock, openrouter, ollama, gateway, sglang
+AI_PROVIDER=google  # or: openai, anthropic, deepseek, siliconflow, doubao, azure, bedrock, openrouter, ollama, gateway, sglang, modelscope
 ```
 
+## Server-Side Multi-Model Configuration
+
+Administrators can configure multiple server-side models that are available to all users without requiring personal API keys.
+
+### Configuration Methods
+
+**Option 1: Environment Variable** (recommended for cloud deployments)
+
+Set `AI_MODELS_CONFIG` as a JSON string:
+
+```bash
+AI_MODELS_CONFIG='{"providers":[{"name":"OpenAI","provider":"openai","models":["gpt-4o"],"default":true}]}'
+```
+
+**Option 2: Config File**
+
+Create an `ai-models.json` file in the project root (or set `AI_MODELS_CONFIG_PATH` to a custom location).
+
+### Example Configuration
+
+```json
+{
+  "providers": [
+    {
+      "name": "OpenAI Production",
+      "provider": "openai",
+      "models": ["gpt-4o", "gpt-4o-mini"],
+      "default": true
+    },
+    {
+      "name": "Custom DeepSeek",
+      "provider": "deepseek",
+      "models": ["deepseek-chat"],
+      "apiKeyEnv": "MY_DEEPSEEK_KEY",
+      "baseUrlEnv": "MY_DEEPSEEK_URL"
+    }
+  ]
+}
+```
+
+### Field Reference
+
+| Field | Required | Description |
+|-------|----------|-------------|
+| `name` | Yes | Display name (supports multiple configs for same provider) |
+| `provider` | Yes | Provider type (`openai`, `anthropic`, `google`, `bedrock`, etc.) |
+| `models` | Yes | List of model IDs |
+| `default` | No | Set to `true` to auto-select this provider's first model as default |
+| `apiKeyEnv` | No | Custom API key env var name (defaults to provider's standard var like `OPENAI_API_KEY`) |
+| `baseUrlEnv` | No | Custom base URL env var name |
+
+### Notes
+
+- API keys and credentials are provided via environment variables. By default, standard var names are used (e.g., `OPENAI_API_KEY`), but you can specify custom var names with `apiKeyEnv`.
+- The `name` field allows multiple configurations for the same provider (e.g., "OpenAI Production" and "OpenAI Staging" both using `provider: "openai"` but with different `apiKeyEnv` values).
+- If config is not present, the app falls back to `AI_PROVIDER`/`AI_MODEL` environment variable configuration.
+
 ## Model Capability Requirements
 
 This task requires exceptionally strong model capabilities, as it involves generating long-form text with strict formatting constraints (draw.io XML).

docs/en/docker.md

@@ -22,6 +22,27 @@ cp env.example .env
 docker run -d -p 3000:3000 --env-file .env ghcr.io/dayuanjiang/next-ai-draw-io:latest
 ```
 
+### Using server-side model configuration
+
+You can mount an `ai-models.json` file into the container to provide multiple server-side models without exposing user API keys:
+
+```bash
+docker run -d -p 3000:3000 \
+  -e OPENAI_API_KEY=your_api_key \
+  -v $(pwd)/ai-models.json:/app/ai-models.json:ro \
+  ghcr.io/dayuanjiang/next-ai-draw-io:latest
+```
+
+If you prefer to keep the config in a different path inside the container, set `AI_MODELS_CONFIG_PATH`:
+
+```bash
+docker run -d -p 3000:3000 \
+  -e OPENAI_API_KEY=your_api_key \
+  -e AI_MODELS_CONFIG_PATH=/config/ai-models.json \
+  -v $(pwd)/ai-models.json:/config/ai-models.json:ro \
+  ghcr.io/dayuanjiang/next-ai-draw-io:latest
+```
+
 Open [http://localhost:3000](http://localhost:3000) in your browser.
 
 Replace the environment variables with your preferred AI provider configuration. See [AI Providers](./ai-providers.md) for available options.

docs/ja/FAQ.md

@@ -0,0 +1,78 @@
+# よくある質問 (FAQ)
+
+---
+
+## 1. PDFをエクスポートできない
+
+**問題**: Web版でPDFエクスポートをクリックすると `convert.diagrams.net/node/export` にリダイレクトされ、その後何も起こらない
+
+**原因**: 埋め込みDraw.ioは直接PDFエクスポートをサポートしておらず、外部変換サービスに依存しているが、iframe内では正常に動作しない
+
+**解決策**: まず画像（PNG）としてエクスポートし、その後PDFに印刷する
+
+**関連Issue**: #539, #125
+
+---
+
+## 2. embed.diagrams.netにアクセスできない（オフライン/イントラネットデプロイ）
+
+**問題**: イントラネット環境で「embed.diagrams.netのサーバーIPアドレスが見つかりません」と表示される
+
+**重要**: `NEXT_PUBLIC_*` 環境変数は**ビルド時**変数であり、JSコードにバンドルされます。**実行時の設定は無効です！**
+
+**解決策**: ビルド時に `args` で渡す必要があります：
+
+```yaml
+# docker-compose.yml
+services:
+  drawio:
+    image: jgraph/drawio:latest
+    ports: ["8080:8080"]
+  next-ai-draw-io:
+    build:
+      context: .
+      args:
+        - NEXT_PUBLIC_DRAWIO_BASE_URL=http://あなたのサーバーIP:8080/
+    ports: ["3000:3000"]
+    env_file: .env
+```
+
+**イントラネットユーザー**: 外部ネットワークでDockerfileを修正してイメージをビルドし、イントラネットに転送する
+
+**関連Issue**: #295, #317
+
+---
+
+## 3. 自前モデルが思考するだけで描画しない
+
+**問題**: ローカルデプロイのモデル（Qwen、LiteLLMなど）が思考過程のみを出力し、図表を生成しない
+
+**考えられる原因**:
+1. **モデルが小さすぎる** - 小さいモデルはtool calling指示に正しく従うことが難しい、32B+パラメータのモデルを推奨
+2. **tool callingが有効になっていない** - モデルサービスでtool use機能を設定する必要がある
+
+**解決策**: tool callingを有効にする、例えばvLLM：
+```bash
+python -m vllm.entrypoints.openai.api_server \
+    --model Qwen/Qwen3-32B \
+    --enable-auto-tool-choice \
+    --tool-call-parser hermes
+```
+
+**関連Issue**: #269, #75
+
+---
+
+## 4. 画像アップロード後「画像が提供されていません」と表示される
+
+**問題**: 画像をアップロードした後、「画像が提供されていません」というエラーが表示される
+
+**考えられる原因**:
+1. モデルがビジョン機能をサポートしていない（Kimi K2、DeepSeek、Qwenテキストモデルなど）
+
+**解決策**:
+- ビジョン対応モデルを使用：GPT-5.2、Claude 4.5 Sonnet、Gemini 3 Pro
+- モデル名に `vision` または `vl` が含まれているものは画像をサポート
+- 最新バージョン（v0.4.9+）にアップデート
+
+**関連Issue**: #324, #421, #469

docs/ja/README_JA.md

@@ -37,11 +37,12 @@ https://github.com/user-attachments/assets/b2eef5f3-b335-4e71-a755-dc2e80931979
     - [インストール](#インストール)
   - [デプロイ](#デプロイ)
     - [EdgeOne Pagesへのデプロイ](#edgeone-pagesへのデプロイ)
-    - [Vercelへのデプロイ（推奨）](#vercelへのデプロイ推奨)
+    - [Vercelへのデプロイ](#vercelへのデプロイ)
     - [Cloudflare Workersへのデプロイ](#cloudflare-workersへのデプロイ)
   - [マルチプロバイダーサポート](#マルチプロバイダーサポート)
   - [仕組み](#仕組み)
   - [サポート＆お問い合わせ](#サポートお問い合わせ)
+  - [よくある質問](#よくある質問)
   - [スター履歴](#スター履歴)
 
 ## 例
@@ -180,7 +181,7 @@ npm run dev
 
 また、Tencent EdgeOne Pagesでデプロイすると、[DeepSeekモデルの毎日の無料クォータ](https://pages.edgeone.ai/document/edge-ai)が付与されます。
 
-### Vercelへのデプロイ（推奨）
+### Vercelへのデプロイ
 
 [![Deploy with Vercel](https://vercel.com/button)](https://vercel.com/new/clone?repository-url=https%3A%2F%2Fgithub.com%2FDayuanJiang%2Fnext-ai-draw-io)
 
@@ -200,11 +201,13 @@ Next.jsアプリをデプロイする最も簡単な方法は、Next.jsの作成
 -   OpenAI
 -   Anthropic
 -   Google AI
+-   Google Vertex AI
 -   Azure OpenAI
 -   Ollama
 -   OpenRouter
 -   DeepSeek
 -   SiliconFlow
+-   ModelScope
 -   SGLang
 -   Vercel AI Gateway
 
@@ -212,6 +215,10 @@ AWS BedrockとOpenRouter以外のすべてのプロバイダーはカスタム
 
 📖 **[詳細なプロバイダー設定ガイド](./ai-providers.md)** - 各プロバイダーの設定手順をご覧ください。
 
+### サーバーサイドマルチモデル設定
+
+管理者は、ユーザーが個人のAPIキーを提供することなく利用できる複数のサーバーサイドモデルを設定できます。`AI_MODELS_CONFIG` 環境変数（JSON文字列）または `ai-models.json` ファイルで設定します。
+
 **モデル要件**：このタスクは厳密なフォーマット制約（draw.io XML）を持つ長文テキスト生成を伴うため、強力なモデル機能が必要です。Claude Sonnet 4.5、GPT-5.1、Gemini 3 Pro、DeepSeek V3.2/R1を推奨します。
 
 注：`claude`シリーズはAWS、Azure、GCPなどのクラウドアーキテクチャロゴ付きのdraw.ioダイアグラムで学習されているため、クラウドアーキテクチャダイアグラムを作成したい場合は最適な選択です。
@@ -238,6 +245,10 @@ AWS BedrockとOpenRouter以外のすべてのプロバイダーはカスタム
 
 -   メール：me[at]jiang.jp
 
+## よくある質問
+
+一般的な問題と解決策については [FAQ](./FAQ.md) をご覧ください。
+
 ## スター履歴
 
 [![Star History Chart](https://api.star-history.com/svg?repos=DayuanJiang/next-ai-draw-io&type=date&legend=top-left)](https://www.star-history.com/#DayuanJiang/next-ai-draw-io&type=date&legend=top-left)

docs/ja/ai-providers.md

@@ -158,6 +158,19 @@ AI_MODEL=llama3.2
 OLLAMA_BASE_URL=http://localhost:11434
 ```
 
+### ModelScope
+
+```bash
+MODELSCOPE_API_KEY=your_api_key
+AI_MODEL=Qwen/Qwen3-235B-A22B-Instruct-2507
+```
+
+任意のカスタムエンドポイント:
+
+```bash
+MODELSCOPE_BASE_URL=https://your-custom-endpoint
+```
+
 ### Vercel AI Gateway
 
 Vercel AI Gateway は、単一の API キーで複数の AI プロバイダーへの統合アクセスを提供します。これにより認証が簡素化され、複数の API キーを管理することなくプロバイダーを切り替えることができます。
@@ -204,6 +217,63 @@ AI_MODEL=openai/gpt-4o
 AI_PROVIDER=google  # または: openai, anthropic, deepseek, siliconflow, doubao, azure, bedrock, openrouter, ollama, gateway, sglang
 ```
 
+## サーバーサイドマルチモデル設定
+
+管理者は、ユーザーが個人のAPIキーを提供することなく利用できる複数のサーバーサイドモデルを設定できます。
+
+### 設定方法
+
+**方法1：環境変数**（クラウドデプロイ推奨）
+
+`AI_MODELS_CONFIG` をJSON文字列として設定：
+
+```bash
+AI_MODELS_CONFIG='{"providers":[{"name":"OpenAI","provider":"openai","models":["gpt-4o"],"default":true}]}'
+```
+
+**方法2：設定ファイル**
+
+プロジェクトルートに `ai-models.json` ファイルを作成します（または `AI_MODELS_CONFIG_PATH` でパスを指定）。
+
+### 設定例
+
+```json
+{
+  "providers": [
+    {
+      "name": "OpenAI Production",
+      "provider": "openai",
+      "models": ["gpt-4o", "gpt-4o-mini"],
+      "default": true
+    },
+    {
+      "name": "Custom DeepSeek",
+      "provider": "deepseek",
+      "models": ["deepseek-chat"],
+      "apiKeyEnv": "MY_DEEPSEEK_KEY",
+      "baseUrlEnv": "MY_DEEPSEEK_URL"
+    }
+  ]
+}
+```
+
+### フィールド説明
+
+| フィールド | 必須 | 説明 |
+|------------|------|------|
+| `name` | はい | 表示名（同一プロバイダーの複数設定をサポート） |
+| `provider` | はい | プロバイダータイプ（`openai`, `anthropic`, `google`, `bedrock` など） |
+| `models` | はい | モデルIDのリスト |
+| `default` | いいえ | `true` に設定すると、そのプロバイダーの最初のモデルがデフォルトで選択されます |
+| `apiKeyEnv` | いいえ | カスタムAPIキー環境変数名（デフォルトは `OPENAI_API_KEY` などの標準変数） |
+| `baseUrlEnv` | いいえ | カスタムBase URL環境変数名 |
+
+### 備考
+
+- APIキーと認証情報は環境変数で提供します。デフォルトは標準変数名（例：`OPENAI_API_KEY`）を使用しますが、`apiKeyEnv` でカスタム変数名を指定できます。
+- `name` フィールドにより同一プロバイダーの複数設定が可能です（例：「OpenAI Production」と「OpenAI Staging」が両方とも `provider: "openai"` を使用しつつ、異なる `apiKeyEnv` を持つ）。
+- 設定が存在しない場合、アプリは `AI_PROVIDER`/`AI_MODEL` 環境変数設定にフォールバックします。
+
 ## モデル性能要件
 
 このタスクは、厳密なフォーマット制約（draw.io XML）を伴う長文テキストの生成を含むため、非常に強力なモデル性能が必要です。

electron/electron.d.ts

@@ -25,6 +25,19 @@ interface ApplyPresetResult {
     env?: Record<string, string>
 }
 
+/** Proxy configuration interface */
+interface ProxyConfig {
+    httpProxy?: string
+    httpsProxy?: string
+}
+
+/** Result of setting proxy */
+interface SetProxyResult {
+    success: boolean
+    error?: string
+    devMode?: boolean
+}
+
 declare global {
     interface Window {
         /** Main window Electron API */
@@ -45,6 +58,10 @@ declare global {
             openFile: () => Promise<string | null>
             /** Save data to file via save dialog */
             saveFile: (data: string) => Promise<boolean>
+            /** Get proxy configuration */
+            getProxy: () => Promise<ProxyConfig>
+            /** Set proxy configuration (saves and restarts server) */
+            setProxy: (config: ProxyConfig) => Promise<SetProxyResult>
         }
 
         /** Settings window Electron API */
@@ -71,4 +88,4 @@ declare global {
     }
 }
 
-export { ConfigPreset, ApplyPresetResult }
+export { ConfigPreset, ApplyPresetResult, ProxyConfig, SetProxyResult }

electron/main/config-manager.ts

@@ -351,6 +351,10 @@ const PROVIDER_ENV_MAP: Record<string, { apiKey: string; baseUrl: string }> = {
         apiKey: "SILICONFLOW_API_KEY",
         baseUrl: "SILICONFLOW_BASE_URL",
     },
+    modelscope: {
+        apiKey: "MODELSCOPE_API_KEY",
+        baseUrl: "MODELSCOPE_BASE_URL",
+    },
     gateway: { apiKey: "AI_GATEWAY_API_KEY", baseUrl: "AI_GATEWAY_BASE_URL" },
     // bedrock and ollama don't use API keys in the same way
     bedrock: { apiKey: "", baseUrl: "" },

electron/main/index.ts

@@ -4,6 +4,7 @@ import { getCurrentPresetEnv } from "./config-manager"
 import { loadEnvFile } from "./env-loader"
 import { registerIpcHandlers } from "./ipc-handlers"
 import { startNextServer, stopNextServer } from "./next-server"
+import { applyProxyToEnv } from "./proxy-manager"
 import { registerSettingsWindowHandlers } from "./settings-window"
 import { createWindow, getMainWindow } from "./window-manager"
 
@@ -24,6 +25,9 @@ if (!gotTheLock) {
     // Load environment variables from .env files
     loadEnvFile()
 
+    // Apply proxy settings from saved config
+    applyProxyToEnv()
+
     // Apply saved preset environment variables (overrides .env)
     const presetEnv = getCurrentPresetEnv()
     for (const [key, value] of Object.entries(presetEnv)) {

electron/main/ipc-handlers.ts

@@ -11,6 +11,12 @@ import {
     updatePreset,
 } from "./config-manager"
 import { restartNextServer } from "./next-server"
+import {
+    applyProxyToEnv,
+    getProxyConfig,
+    type ProxyConfig,
+    saveProxyConfig,
+} from "./proxy-manager"
 
 /**
  * Allowed configuration keys for presets
@@ -209,4 +215,40 @@ export function registerIpcHandlers(): void {
             return setCurrentPreset(id)
         },
     )
+
+    // ==================== Proxy Settings ====================
+
+    ipcMain.handle("get-proxy", () => {
+        return getProxyConfig()
+    })
+
+    ipcMain.handle("set-proxy", async (_event, config: ProxyConfig) => {
+        try {
+            // Save config to file
+            saveProxyConfig(config)
+
+            // Apply to current process environment
+            applyProxyToEnv()
+
+            const isDev = process.env.NODE_ENV === "development"
+
+            if (isDev) {
+                // In development, env vars are already applied
+                // Next.js dev server may need manual restart
+                return { success: true, devMode: true }
+            }
+
+            // Production: restart Next.js server to pick up new env vars
+            await restartNextServer()
+            return { success: true }
+        } catch (error) {
+            return {
+                success: false,
+                error:
+                    error instanceof Error
+                        ? error.message
+                        : "Failed to apply proxy settings",
+            }
+        }
+    })
 }

electron/main/next-server.ts

@@ -69,6 +69,8 @@ export async function startNextServer(): Promise<string> {
         NODE_ENV: "production",
         PORT: String(port),
         HOSTNAME: "localhost",
+        // Enable Node.js built-in proxy support for fetch (Node.js 24+)
+        NODE_USE_ENV_PROXY: "1",
     }
 
     // Set cache directory to a writable location (user's app data folder)
@@ -85,6 +87,13 @@ export async function startNextServer(): Promise<string> {
         }
     }
 
+    // Debug: log proxy-related env vars
+    console.log("Proxy env vars being passed to server:", {
+        HTTP_PROXY: env.HTTP_PROXY || env.http_proxy || "not set",
+        HTTPS_PROXY: env.HTTPS_PROXY || env.https_proxy || "not set",
+        NODE_USE_ENV_PROXY: env.NODE_USE_ENV_PROXY || "not set",
+    })
+
     // Use Electron's utilityProcess API for running Node.js in background
     // This is the recommended way to run Node.js code in Electron
     serverProcess = utilityProcess.fork(serverPath, [], {
@@ -114,13 +123,41 @@ export async function startNextServer(): Promise<string> {
 }
 
 /**
- * Stop the Next.js server process
+ * Stop the Next.js server process and wait for it to exit
  */
-export function stopNextServer(): void {
+export async function stopNextServer(): Promise<void> {
     if (serverProcess) {
         console.log("Stopping Next.js server...")
+
+        // Create a promise that resolves when the process exits
+        const exitPromise = new Promise<void>((resolve) => {
+            const proc = serverProcess
+            if (!proc) {
+                resolve()
+                return
+            }
+
+            const onExit = () => {
+                resolve()
+            }
+
+            proc.once("exit", onExit)
+
+            // Timeout after 5 seconds
+            setTimeout(() => {
+                proc.removeListener("exit", onExit)
+                resolve()
+            }, 5000)
+        })
+
         serverProcess.kill()
         serverProcess = null
+
+        // Wait for process to exit
+        await exitPromise
+
+        // Additional wait for OS to release port
+        await new Promise((resolve) => setTimeout(resolve, 500))
     }
 }
 
@@ -150,8 +187,8 @@ async function waitForServerStop(timeout = 5000): Promise<void> {
 export async function restartNextServer(): Promise<string> {
     console.log("Restarting Next.js server...")
 
-    // Stop the current server
-    stopNextServer()
+    // Stop the current server and wait for it to exit
+    await stopNextServer()
 
     // Wait for the port to be released
     await waitForServerStop()

electron/main/proxy-manager.ts

@@ -0,0 +1,75 @@
+import { app } from "electron"
+import * as fs from "fs"
+import * as path from "path"
+import type { ProxyConfig } from "../electron.d"
+
+export type { ProxyConfig }
+
+const CONFIG_FILE = "proxy-config.json"
+
+function getConfigPath(): string {
+    return path.join(app.getPath("userData"), CONFIG_FILE)
+}
+
+/**
+ * Load proxy configuration from JSON file
+ */
+export function loadProxyConfig(): ProxyConfig {
+    try {
+        const configPath = getConfigPath()
+        if (fs.existsSync(configPath)) {
+            const data = fs.readFileSync(configPath, "utf-8")
+            return JSON.parse(data) as ProxyConfig
+        }
+    } catch (error) {
+        console.error("Failed to load proxy config:", error)
+    }
+    return {}
+}
+
+/**
+ * Save proxy configuration to JSON file
+ */
+export function saveProxyConfig(config: ProxyConfig): void {
+    try {
+        const configPath = getConfigPath()
+        fs.writeFileSync(configPath, JSON.stringify(config, null, 2), "utf-8")
+    } catch (error) {
+        console.error("Failed to save proxy config:", error)
+        throw error
+    }
+}
+
+/**
+ * Apply proxy configuration to process.env
+ * Must be called BEFORE starting the Next.js server
+ */
+export function applyProxyToEnv(): void {
+    const config = loadProxyConfig()
+
+    if (config.httpProxy) {
+        process.env.HTTP_PROXY = config.httpProxy
+        process.env.http_proxy = config.httpProxy
+    } else {
+        delete process.env.HTTP_PROXY
+        delete process.env.http_proxy
+    }
+
+    if (config.httpsProxy) {
+        process.env.HTTPS_PROXY = config.httpsProxy
+        process.env.https_proxy = config.httpsProxy
+    } else {
+        delete process.env.HTTPS_PROXY
+        delete process.env.https_proxy
+    }
+}
+
+/**
+ * Get current proxy configuration (from process.env)
+ */
+export function getProxyConfig(): ProxyConfig {
+    return {
+        httpProxy: process.env.HTTP_PROXY || process.env.http_proxy || "",
+        httpsProxy: process.env.HTTPS_PROXY || process.env.https_proxy || "",
+    }
+}

electron/preload/index.ts

@@ -21,4 +21,9 @@ contextBridge.exposeInMainWorld("electronAPI", {
     // File operations
     openFile: () => ipcRenderer.invoke("dialog-open-file"),
     saveFile: (data: string) => ipcRenderer.invoke("dialog-save-file", data),
+
+    // Proxy settings
+    getProxy: () => ipcRenderer.invoke("get-proxy"),
+    setProxy: (config: { httpProxy?: string; httpsProxy?: string }) =>
+        ipcRenderer.invoke("set-proxy", config),
 })

electron/settings/index.html

@@ -55,6 +55,7 @@
                             <option value="openrouter">OpenRouter</option>
                             <option value="deepseek">DeepSeek</option>
                             <option value="siliconflow">SiliconFlow</option>
+                            <option value="modelscope">ModelScope</option>
                             <option value="ollama">Ollama (Local)</option>
                         </select>
                     </div>

electron/settings/settings.js

@@ -288,6 +288,7 @@ function getProviderLabel(provider) {
         openrouter: "OpenRouter",
         deepseek: "DeepSeek",
         siliconflow: "SiliconFlow",
+        modelscope: "ModelScope",
         ollama: "Ollama",
     }
     return labels[provider] || provider

env.example

@@ -1,6 +1,6 @@
 # AI Provider Configuration
 # AI_PROVIDER: Which provider to use
-# Options: bedrock, openai, anthropic, google, azure, ollama, openrouter, deepseek, siliconflow, gateway
+# Options: bedrock, openai, anthropic, google, vertexai, azure, ollama, openrouter, deepseek, siliconflow, gateway
 # Default: bedrock
 AI_PROVIDER=bedrock
 
@@ -40,6 +40,14 @@ AI_MODEL=global.anthropic.claude-sonnet-4-5-20250929-v1:0
 # GOOGLE_THINKING_BUDGET=8192                # Optional: Gemini 2.5 thinking budget in tokens (for more/less thinking)
 # GOOGLE_THINKING_LEVEL=high                 # Optional: Gemini 3 thinking level (low/high)
 
+# Google Vertex AI Configuration (Enterprise GCP)
+# For enterprise users needing data residency, VPC Service Controls, or GCP integration
+# GOOGLE_VERTEX_API_KEY=                          # Required: Express Mode API key
+# GOOGLE_VERTEX_BASE_URL=https://...              # Optional: Custom endpoint URL
+# Note: Gemini 2.5/3 models automatically enable reasoning display (includeThoughts: true)
+# GOOGLE_VERTEX_THINKING_BUDGET=8192              # Optional: Gemini 2.5 thinking budget in tokens (1024-100000)
+# GOOGLE_VERTEX_THINKING_LEVEL=high               # Optional: Gemini 3 thinking level (minimal/low/medium/high)
+
 # Azure OpenAI Configuration
 # Configure endpoint using ONE of these methods:
 #   1. AZURE_RESOURCE_NAME - SDK constructs: https://{name}.openai.azure.com/openai/v1{path}
@@ -72,6 +80,10 @@ AI_MODEL=global.anthropic.claude-sonnet-4-5-20250929-v1:0
 # SGLANG_API_KEY=your-sglang-api-key
 # SGLANG_BASE_URL=http://127.0.0.1:8000/v1  # Your SGLang endpoint
 
+# ModelScope Configuration
+# MODELSCOPE_API_KEY=ms-...
+# MODELSCOPE_BASE_URL=https://api-inference.modelscope.cn/v1  # Optional: Custom endpoint
+
 # ByteDance Doubao Configuration (via Volcengine)
 # DOUBAO_API_KEY=your-doubao-api-key
 # DOUBAO_BASE_URL=https://ark.cn-beijing.volces.com/api/v3  # ByteDance Volcengine endpoint
@@ -89,6 +101,11 @@ AI_MODEL=global.anthropic.claude-sonnet-4-5-20250929-v1:0
 # LANGFUSE_SECRET_KEY=sk-lf-...
 # LANGFUSE_BASEURL=https://cloud.langfuse.com  # EU region, use https://us.cloud.langfuse.com for US
 
+# Optional server-side multi-model configuration
+# If set, points to a JSON file with server-provided models (see README for schema).
+# Default: ./ai-models.json in project root
+# AI_MODELS_CONFIG_PATH=/path/to/ai-models.json
+
 # Temperature (Optional)
 # Controls randomness in AI responses. Lower = more deterministic.
 # Leave unset for models that don't support temperature (e.g., GPT-5.1 reasoning models)

hooks/use-diagram-tool-handlers.ts

@@ -1,4 +1,5 @@
 import type { MutableRefObject } from "react"
+import type { DiagramOperation } from "@/components/chat/types"
 import { isMxCellXmlComplete, wrapWithMxFile } from "@/lib/utils"
 
 const DEBUG = process.env.NODE_ENV === "development"
@@ -29,12 +30,6 @@ type AddToolOutputParams = AddToolOutputSuccess | AddToolOutputError
 
 type AddToolOutputFn = (params: AddToolOutputParams) => void
 
-interface DiagramOperation {
-    operation: "update" | "add" | "delete"
-    cell_id: string
-    new_xml?: string
-}
-
 interface UseDiagramToolHandlersParams {
     partialXmlRef: MutableRefObject<string>
     editDiagramOriginalXmlRef: MutableRefObject<Map<string, string>>

hooks/use-model-config.ts

@@ -1,6 +1,7 @@
 "use client"
 
 import { useCallback, useEffect, useState } from "react"
+import type { FlattenedServerModel } from "@/lib/server-model-config"
 import { STORAGE_KEYS } from "@/lib/storage"
 import {
     createEmptyConfig,
@@ -132,14 +133,56 @@ export interface UseModelConfigReturn {
 export function useModelConfig(): UseModelConfigReturn {
     const [config, setConfig] = useState<MultiModelConfig>(createEmptyConfig)
     const [isLoaded, setIsLoaded] = useState(false)
+    const [serverModels, setServerModels] = useState<FlattenedServerModel[]>([])
+    const [serverLoaded, setServerLoaded] = useState(false)
 
-    // Load config on mount
+    // Load client config on mount
     useEffect(() => {
         const loaded = loadConfig()
         setConfig(loaded)
         setIsLoaded(true)
     }, [])
 
+    // Load server models on mount (if any)
+    useEffect(() => {
+        if (typeof window === "undefined") return
+
+        fetch("/api/server-models")
+            .then((res) => {
+                if (!res.ok) {
+                    console.error(
+                        "Failed to load server models:",
+                        res.status,
+                        res.statusText,
+                    )
+                    throw new Error(`Request failed with status ${res.status}`)
+                }
+                return res.json()
+            })
+            .then((data) => {
+                const raw: FlattenedServerModel[] = data?.models || []
+                setServerModels(raw)
+                setServerLoaded(true)
+
+                // Auto-select default server model if no model is currently selected
+                setConfig((prev) => {
+                    if (!prev.selectedModelId && raw.length > 0) {
+                        const defaultModel = raw.find((m) => m.isDefault)
+                        if (defaultModel) {
+                            return { ...prev, selectedModelId: defaultModel.id }
+                        }
+                        // If no default marked, use first server model
+                        return { ...prev, selectedModelId: raw[0].id }
+                    }
+                    return prev
+                })
+            })
+            .catch((error) => {
+                console.error("Error while loading server models:", error)
+                setServerLoaded(true)
+            })
+    }, [])
+
     // Save config whenever it changes (after initial load)
     useEffect(() => {
         if (isLoaded) {
@@ -148,9 +191,33 @@ export function useModelConfig(): UseModelConfigReturn {
     }, [config, isLoaded])
 
     // Derived state
-    const models = flattenModels(config)
+    const userModels = flattenModels(config)
+
+    const models: FlattenedModel[] = [
+        // Server models (read-only, credentials from env)
+        ...serverModels.map((m) => ({
+            id: m.id,
+            modelId: m.modelId,
+            provider: m.provider,
+            providerLabel: `Server · ${m.providerLabel}`,
+            apiKey: "",
+            baseUrl: undefined,
+            awsAccessKeyId: undefined,
+            awsSecretAccessKey: undefined,
+            awsRegion: undefined,
+            awsSessionToken: undefined,
+            validated: true,
+            source: "server" as const,
+            isDefault: m.isDefault,
+            apiKeyEnv: m.apiKeyEnv,
+            baseUrlEnv: m.baseUrlEnv,
+        })),
+        // User models from local configuration
+        ...userModels,
+    ]
+
     const selectedModel = config.selectedModelId
-        ? findModelById(config, config.selectedModelId)
+        ? models.find((m) => m.id === config.selectedModelId)
         : undefined
 
     // Actions
@@ -282,7 +349,7 @@ export function useModelConfig(): UseModelConfigReturn {
 
     return {
         config,
-        isLoaded,
+        isLoaded: isLoaded && serverLoaded,
         models,
         selectedModel,
         selectedModelId: config.selectedModelId,
@@ -314,6 +381,10 @@ export function getSelectedAIConfig(): {
     awsSecretAccessKey: string
     awsRegion: string
     awsSessionToken: string
+    // Selected model ID (for server model lookup)
+    selectedModelId: string
+    // Vertex AI credentials (Express Mode)
+    vertexApiKey: string
 } {
     const empty = {
         accessCode: "",
@@ -325,6 +396,8 @@ export function getSelectedAIConfig(): {
         awsSecretAccessKey: "",
         awsRegion: "",
         awsSessionToken: "",
+        selectedModelId: "",
+        vertexApiKey: "",
     }
 
     if (typeof window === "undefined") return empty
@@ -347,6 +420,8 @@ export function getSelectedAIConfig(): {
             awsSecretAccessKey: "",
             awsRegion: "",
             awsSessionToken: "",
+            selectedModelId: "",
+            vertexApiKey: "",
         }
     }
 
@@ -357,12 +432,32 @@ export function getSelectedAIConfig(): {
         return { ...empty, accessCode }
     }
 
-    // No selected model = use server default
+    // No selected model = use server default (AI_PROVIDER/AI_MODEL/env auto-detect)
     if (!config.selectedModelId) {
         return { ...empty, accessCode }
     }
 
-    // Find selected model
+    // Server-side model selection (id = "server:<name-slug>:<modelId>")
+    // Provider is resolved server-side via findServerModelById()
+    if (config.selectedModelId.startsWith("server:")) {
+        const parts = config.selectedModelId.split(":")
+        const nameSlug = parts[1] || ""
+        const modelId = parts.slice(2).join(":") // Preserve Bedrock-style IDs
+
+        return {
+            ...empty,
+            accessCode,
+            // Note: nameSlug is NOT the provider, but we send it for backwards compat
+            // Server uses selectedModelId to lookup the actual provider
+            aiProvider: nameSlug,
+            aiBaseUrl: "",
+            aiApiKey: "",
+            aiModel: modelId,
+            selectedModelId: config.selectedModelId,
+        }
+    }
+
+    // Find selected user-defined model
     const model = findModelById(config, config.selectedModelId)
     if (!model) {
         return { ...empty, accessCode }
@@ -379,5 +474,8 @@ export function getSelectedAIConfig(): {
         awsSecretAccessKey: model.awsSecretAccessKey || "",
         awsRegion: model.awsRegion || "",
         awsSessionToken: model.awsSessionToken || "",
+        selectedModelId: config.selectedModelId || "",
+        // Vertex AI credentials (Express Mode)
+        vertexApiKey: model.vertexApiKey || "",
     }
 }

lib/ai-config.ts

@@ -1,26 +0,0 @@
-import { STORAGE_KEYS } from "./storage"
-
-/**
- * Get AI configuration from localStorage.
- * Returns API keys and settings for custom AI providers.
- * Used to override server defaults when user provides their own API key.
- */
-export function getAIConfig() {
-    if (typeof window === "undefined") {
-        return {
-            accessCode: "",
-            aiProvider: "",
-            aiBaseUrl: "",
-            aiApiKey: "",
-            aiModel: "",
-        }
-    }
-
-    return {
-        accessCode: localStorage.getItem(STORAGE_KEYS.accessCode) || "",
-        aiProvider: localStorage.getItem(STORAGE_KEYS.aiProvider) || "",
-        aiBaseUrl: localStorage.getItem(STORAGE_KEYS.aiBaseUrl) || "",
-        aiApiKey: localStorage.getItem(STORAGE_KEYS.aiApiKey) || "",
-        aiModel: localStorage.getItem(STORAGE_KEYS.aiModel) || "",
-    }
-}

lib/ai-providers.ts

@@ -4,25 +4,14 @@ import { azure, createAzure } from "@ai-sdk/azure"
 import { createDeepSeek, deepseek } from "@ai-sdk/deepseek"
 import { createGateway, gateway } from "@ai-sdk/gateway"
 import { createGoogleGenerativeAI, google } from "@ai-sdk/google"
+import { createVertex } from "@ai-sdk/google-vertex"
 import { createOpenAI, openai } from "@ai-sdk/openai"
 import { fromNodeProviderChain } from "@aws-sdk/credential-providers"
 import { createOpenRouter } from "@openrouter/ai-sdk-provider"
 import { createOllama, ollama } from "ollama-ai-provider-v2"
+import type { ProviderName } from "@/lib/types/model-config"
 
-export type ProviderName =
-    | "bedrock"
-    | "openai"
-    | "anthropic"
-    | "google"
-    | "azure"
-    | "ollama"
-    | "openrouter"
-    | "deepseek"
-    | "siliconflow"
-    | "sglang"
-    | "gateway"
-    | "edgeone"
-    | "doubao"
+export type { ProviderName }
 
 interface ModelConfig {
     model: any
@@ -41,8 +30,13 @@ export interface ClientOverrides {
     awsSecretAccessKey?: string | null
     awsRegion?: string | null
     awsSessionToken?: string | null
+    // Vertex AI config
+    vertexApiKey?: string | null // Express Mode API key
     // Custom headers (e.g., for EdgeOne cookie auth)
     headers?: Record<string, string>
+    // Custom env var names for server models (allows multiple API keys per provider)
+    apiKeyEnv?: string
+    baseUrlEnv?: string
 }
 
 // Providers that can be used with client-provided API keys
@@ -50,6 +44,7 @@ const ALLOWED_CLIENT_PROVIDERS: ProviderName[] = [
     "openai",
     "anthropic",
     "google",
+    "vertexai",
     "azure",
     "bedrock",
     "openrouter",
@@ -59,6 +54,7 @@ const ALLOWED_CLIENT_PROVIDERS: ProviderName[] = [
     "gateway",
     "edgeone",
     "doubao",
+    "modelscope",
 ]
 
 // Bedrock provider options for Anthropic beta features
@@ -73,6 +69,62 @@ const ANTHROPIC_BETA_HEADERS = {
     "anthropic-beta": "fine-grained-tool-streaming-2025-05-14",
 }
 
+/**
+ * Resolve baseURL based on whether user is providing their own API key.
+ * When user provides their own API key, we should NOT fall back to server's
+ * baseURL environment variable - user credentials should only be sent to
+ * user-specified endpoints or official provider endpoints.
+ *
+ * @param userApiKey - User-provided API key (if any)
+ * @param userBaseUrl - User-provided base URL (if any)
+ * @param serverBaseUrl - Server's base URL from environment variable
+ * @param defaultBaseUrl - Provider's official/default base URL (optional)
+ * @returns The resolved base URL to use
+ */
+export function resolveBaseURL(
+    userApiKey: string | null | undefined,
+    userBaseUrl: string | null | undefined,
+    serverBaseUrl: string | undefined,
+    defaultBaseUrl?: string,
+): string | undefined {
+    if (userApiKey) {
+        // User provides their own API key - only use user's baseUrl or default
+        return userBaseUrl || defaultBaseUrl || undefined
+    }
+    // No user API key - fall back to server config
+    return userBaseUrl || serverBaseUrl || defaultBaseUrl || undefined
+}
+
+/**
+ * Resolve API key from custom env var name or default env var.
+ * Supports multiple API keys per provider via ai-models.json apiKeyEnv config.
+ *
+ * Priority:
+ * 1. User-provided API key (overrides.apiKey)
+ * 2. Custom env var from ai-models.json (overrides.apiKeyEnv)
+ * 3. Default provider env var (defaultEnvVar)
+ */
+function resolveApiKey(
+    overrides: ClientOverrides | undefined,
+    defaultEnvVar: string,
+): string | undefined {
+    if (overrides?.apiKey) return overrides.apiKey
+    if (overrides?.apiKeyEnv) return process.env[overrides.apiKeyEnv]
+    return process.env[defaultEnvVar]
+}
+
+/**
+ * Resolve base URL from custom env var name or default env var.
+ * Supports multiple base URLs per provider via ai-models.json baseUrlEnv config.
+ */
+function resolveBaseUrlEnv(
+    overrides: ClientOverrides | undefined,
+    defaultEnvVar: string,
+): string | undefined {
+    if (overrides?.baseUrlEnv) return process.env[overrides.baseUrlEnv]
+    return process.env[defaultEnvVar]
+}
+
 /**
  * Safely parse integer from environment variable with validation
  */
@@ -107,6 +159,8 @@ function parseIntSafe(
  * - ANTHROPIC_THINKING_TYPE: Anthropic thinking type (enabled)
  * - GOOGLE_THINKING_BUDGET: Google Gemini 2.5 thinking budget in tokens (1024-100000)
  * - GOOGLE_THINKING_LEVEL: Google Gemini 3 thinking level (low/high)
+ * - GOOGLE_VERTEX_THINKING_BUDGET: Vertex AI Gemini 2.5 thinking budget in tokens (1024-100000)
+ * - GOOGLE_VERTEX_THINKING_LEVEL: Vertex AI Gemini 3 thinking level (low/high)
  * - AZURE_REASONING_EFFORT: Azure/OpenAI reasoning effort (low/medium/high)
  * - AZURE_REASONING_SUMMARY: Azure reasoning summary (none/brief/detailed)
  * - BEDROCK_REASONING_BUDGET_TOKENS: Bedrock Claude reasoning budget in tokens (1024-64000)
@@ -271,7 +325,46 @@ function buildProviderOptions(
             }
             break
         }
+        case "vertexai": {
+            const thinkingBudget = parseIntSafe(
+                process.env.GOOGLE_VERTEX_THINKING_BUDGET,
+                "GOOGLE_VERTEX_THINKING_BUDGET",
+                1024,
+                100000,
+            )
+            const thinkingLevel = process.env.GOOGLE_VERTEX_THINKING_LEVEL
 
+            if (
+                modelId &&
+                (modelId.includes("gemini-2") ||
+                    modelId.includes("gemini-3") ||
+                    modelId.includes("gemini2") ||
+                    modelId.includes("gemini3"))
+            ) {
+                const thinkingConfig: Record<string, any> = {
+                    includeThoughts: true,
+                }
+
+                const isGemini3 =
+                    modelId?.includes("gemini-3") ||
+                    modelId?.includes("gemini3")
+                const isGemini25 =
+                    modelId?.includes("2.5") || modelId?.includes("2-5")
+
+                if (isGemini3 && thinkingLevel) {
+                    // Vertex AI provider in AI SDK supports more granular levels (minimal/low/medium/high)
+                    thinkingConfig.thinkingLevel = thinkingLevel as
+                        | "minimal"
+                        | "low"
+                        | "medium"
+                        | "high"
+                } else if (isGemini25 && thinkingBudget) {
+                    thinkingConfig.thinkingBudget = thinkingBudget
+                }
+                options.google = { thinkingConfig }
+            }
+            break
+        }
         case "azure": {
             const reasoningEffort = process.env.AZURE_REASONING_EFFORT
             const reasoningSummary = process.env.AZURE_REASONING_SUMMARY
@@ -353,6 +446,7 @@ function buildProviderOptions(
         case "siliconflow":
         case "sglang":
         case "gateway":
+        case "modelscope":
         case "doubao": {
             // These providers don't have reasoning configs in AI SDK yet
             // Gateway passes through to underlying providers which handle their own configs
@@ -372,6 +466,7 @@ const PROVIDER_ENV_VARS: Record<ProviderName, string | null> = {
     openai: "OPENAI_API_KEY",
     anthropic: "ANTHROPIC_API_KEY",
     google: "GOOGLE_GENERATIVE_AI_API_KEY",
+    vertexai: "GOOGLE_VERTEX_API_KEY",
     azure: "AZURE_API_KEY",
     ollama: null, // No credentials needed for local Ollama
     openrouter: "OPENROUTER_API_KEY",
@@ -381,6 +476,7 @@ const PROVIDER_ENV_VARS: Record<ProviderName, string | null> = {
     gateway: "AI_GATEWAY_API_KEY",
     edgeone: null, // No credentials needed - uses EdgeOne Edge AI
     doubao: "DOUBAO_API_KEY",
+    modelscope: "MODELSCOPE_API_KEY",
 }
 
 /**
@@ -418,9 +514,15 @@ function detectProvider(): ProviderName | null {
 
 /**
  * Validate that required API keys are present for the selected provider
+ * @param provider - The provider to validate
+ * @param customApiKeyEnv - Optional custom env var name (from ai-models.json apiKeyEnv)
  */
-function validateProviderCredentials(provider: ProviderName): void {
-    const requiredVar = PROVIDER_ENV_VARS[provider]
+function validateProviderCredentials(
+    provider: ProviderName,
+    customApiKeyEnv?: string,
+): void {
+    // Use custom env var name if provided, otherwise use default
+    const requiredVar = customApiKeyEnv || PROVIDER_ENV_VARS[provider]
     if (requiredVar && !process.env[requiredVar]) {
         throw new Error(
             `${requiredVar} environment variable is required for ${provider} provider. ` +
@@ -445,7 +547,7 @@ function validateProviderCredentials(provider: ProviderName): void {
  * Get the AI model based on environment variables
  *
  * Environment variables:
- * - AI_PROVIDER: The provider to use (bedrock, openai, anthropic, google, azure, ollama, openrouter, deepseek, siliconflow, sglang, gateway)
+ * - AI_PROVIDER: The provider to use (bedrock, openai, anthropic, google, azure, ollama, openrouter, deepseek, siliconflow, sglang, gateway, modelscope)
  * - AI_MODEL: The model ID/name for the selected provider
  *
  * Provider-specific env vars:
@@ -460,9 +562,11 @@ function validateProviderCredentials(provider: ProviderName): void {
  * - DEEPSEEK_API_KEY: DeepSeek API key
  * - DEEPSEEK_BASE_URL: DeepSeek endpoint (optional)
  * - SILICONFLOW_API_KEY: SiliconFlow API key
- * - SILICONFLOW_BASE_URL: SiliconFlow endpoint (optional, defaults to https://api.siliconflow.com/v1)
+ * - SILICONFLOW_BASE_URL: SiliconFlow endpoint (optional, defaults to https://api.siliconflow.cn/v1)
  * - SGLANG_API_KEY: SGLang API key
  * - SGLANG_BASE_URL: SGLang endpoint (optional)
+ * - MODELSCOPE_API_KEY: ModelScope API key
+ * - MODELSCOPE_BASE_URL: ModelScope endpoint (optional)
  */
 export function getAIModel(overrides?: ClientOverrides): ModelConfig {
     // SECURITY: Prevent SSRF attacks (GHSA-9qf7-mprq-9qgm)
@@ -472,6 +576,7 @@ export function getAIModel(overrides?: ClientOverrides): ModelConfig {
     if (
         overrides?.baseUrl &&
         !overrides?.apiKey &&
+        !(overrides?.provider === "vertexai" && overrides?.vertexApiKey) &&
         overrides?.provider !== "edgeone"
     ) {
         throw new Error(
@@ -481,7 +586,11 @@ export function getAIModel(overrides?: ClientOverrides): ModelConfig {
     }
 
     // Check if client is providing their own provider override
-    const isClientOverride = !!(overrides?.provider && overrides?.apiKey)
+    const isClientOverride = !!(
+        overrides?.provider &&
+        (overrides?.apiKey ||
+            (overrides?.provider === "vertexai" && overrides?.vertexApiKey))
+    )
 
     // Use client override if provided, otherwise fall back to env vars
     const modelId = overrides?.modelId || process.env.AI_MODEL
@@ -537,6 +646,7 @@ export function getAIModel(overrides?: ClientOverrides): ModelConfig {
                         `- AZURE_API_KEY for Azure\n` +
                         `- SILICONFLOW_API_KEY for SiliconFlow\n` +
                         `- SGLANG_API_KEY for SGLang\n` +
+                        `- MODELSCOPE_API_KEY for ModelScope\n` +
                         `Or set AI_PROVIDER=ollama for local Ollama.`,
                 )
             } else {
@@ -550,7 +660,7 @@ export function getAIModel(overrides?: ClientOverrides): ModelConfig {
 
     // Only validate server credentials if client isn't providing their own API key
     if (!isClientOverride) {
-        validateProviderCredentials(provider)
+        validateProviderCredentials(provider, overrides?.apiKeyEnv)
     }
 
     console.log(`[AI Provider] Initializing ${provider} with model: ${modelId}`)
@@ -600,8 +710,16 @@ export function getAIModel(overrides?: ClientOverrides): ModelConfig {
         }
 
         case "openai": {
-            const apiKey = overrides?.apiKey || process.env.OPENAI_API_KEY
-            const baseURL = overrides?.baseUrl || process.env.OPENAI_BASE_URL
+            const apiKey = resolveApiKey(overrides, "OPENAI_API_KEY")
+            const serverBaseUrl = resolveBaseUrlEnv(
+                overrides,
+                "OPENAI_BASE_URL",
+            )
+            const baseURL = resolveBaseURL(
+                overrides?.apiKey,
+                overrides?.baseUrl,
+                serverBaseUrl,
+            )
             if (baseURL) {
                 // Custom base URL = third-party proxy, use Chat Completions API
                 // for compatibility (most proxies don't support /responses endpoint)
@@ -619,11 +737,17 @@ export function getAIModel(overrides?: ClientOverrides): ModelConfig {
         }
 
         case "anthropic": {
-            const apiKey = overrides?.apiKey || process.env.ANTHROPIC_API_KEY
-            const baseURL =
-                overrides?.baseUrl ||
-                process.env.ANTHROPIC_BASE_URL ||
-                "https://api.anthropic.com/v1"
+            const apiKey = resolveApiKey(overrides, "ANTHROPIC_API_KEY")
+            const serverBaseUrl = resolveBaseUrlEnv(
+                overrides,
+                "ANTHROPIC_BASE_URL",
+            )
+            const baseURL = resolveBaseURL(
+                overrides?.apiKey,
+                overrides?.baseUrl,
+                serverBaseUrl,
+                "https://api.anthropic.com/v1",
+            )
             const customProvider = createAnthropic({
                 apiKey,
                 baseURL,
@@ -636,9 +760,19 @@ export function getAIModel(overrides?: ClientOverrides): ModelConfig {
         }
 
         case "google": {
-            const apiKey =
-                overrides?.apiKey || process.env.GOOGLE_GENERATIVE_AI_API_KEY
-            const baseURL = overrides?.baseUrl || process.env.GOOGLE_BASE_URL
+            const apiKey = resolveApiKey(
+                overrides,
+                "GOOGLE_GENERATIVE_AI_API_KEY",
+            )
+            const serverBaseUrl = resolveBaseUrlEnv(
+                overrides,
+                "GOOGLE_BASE_URL",
+            )
+            const baseURL = resolveBaseURL(
+                overrides?.apiKey,
+                overrides?.baseUrl,
+                serverBaseUrl,
+            )
             if (baseURL || overrides?.apiKey) {
                 const customGoogle = createGoogleGenerativeAI({
                     apiKey,
@@ -650,11 +784,42 @@ export function getAIModel(overrides?: ClientOverrides): ModelConfig {
             }
             break
         }
+        case "vertexai": {
+            // Express Mode: Use API key for authentication
+            const vertexApiKey =
+                overrides?.vertexApiKey || process.env.GOOGLE_VERTEX_API_KEY
+
+            if (!vertexApiKey) {
+                throw new Error(
+                    "Vertex AI requires an API key for Express Mode. " +
+                        "Get one from Google Cloud Console or set GOOGLE_VERTEX_API_KEY environment variable.",
+                )
+            }
+
+            // Support custom base URL from env or client override
+            const baseURL =
+                overrides?.baseUrl || process.env.GOOGLE_VERTEX_BASE_URL
+
+            const vertexProvider = createVertex({
+                apiKey: vertexApiKey,
+                ...(baseURL && { baseURL }),
+            })
+            model = vertexProvider(modelId)
+            break
+        }
 
         case "azure": {
-            const apiKey = overrides?.apiKey || process.env.AZURE_API_KEY
-            const baseURL = overrides?.baseUrl || process.env.AZURE_BASE_URL
-            const resourceName = process.env.AZURE_RESOURCE_NAME
+            const apiKey = resolveApiKey(overrides, "AZURE_API_KEY")
+            const serverBaseUrl = resolveBaseUrlEnv(overrides, "AZURE_BASE_URL")
+            const baseURL = resolveBaseURL(
+                overrides?.apiKey,
+                overrides?.baseUrl,
+                serverBaseUrl,
+            )
+            // Only use server's resourceName if user is NOT providing their own API key
+            const resourceName = overrides?.apiKey
+                ? undefined
+                : process.env.AZURE_RESOURCE_NAME
             // Azure requires either baseURL or resourceName to construct the endpoint
             // resourceName constructs: https://{resourceName}.openai.azure.com/openai/v1{path}
             if (baseURL || resourceName || overrides?.apiKey) {
@@ -683,9 +848,16 @@ export function getAIModel(overrides?: ClientOverrides): ModelConfig {
             break
 
         case "openrouter": {
-            const apiKey = overrides?.apiKey || process.env.OPENROUTER_API_KEY
-            const baseURL =
-                overrides?.baseUrl || process.env.OPENROUTER_BASE_URL
+            const apiKey = resolveApiKey(overrides, "OPENROUTER_API_KEY")
+            const serverBaseUrl = resolveBaseUrlEnv(
+                overrides,
+                "OPENROUTER_BASE_URL",
+            )
+            const baseURL = resolveBaseURL(
+                overrides?.apiKey,
+                overrides?.baseUrl,
+                serverBaseUrl,
+            )
             const openrouter = createOpenRouter({
                 apiKey,
                 ...(baseURL && { baseURL }),
@@ -695,8 +867,16 @@ export function getAIModel(overrides?: ClientOverrides): ModelConfig {
         }
 
         case "deepseek": {
-            const apiKey = overrides?.apiKey || process.env.DEEPSEEK_API_KEY
-            const baseURL = overrides?.baseUrl || process.env.DEEPSEEK_BASE_URL
+            const apiKey = resolveApiKey(overrides, "DEEPSEEK_API_KEY")
+            const serverBaseUrl = resolveBaseUrlEnv(
+                overrides,
+                "DEEPSEEK_BASE_URL",
+            )
+            const baseURL = resolveBaseURL(
+                overrides?.apiKey,
+                overrides?.baseUrl,
+                serverBaseUrl,
+            )
             if (baseURL || overrides?.apiKey) {
                 const customDeepSeek = createDeepSeek({
                     apiKey,
@@ -710,11 +890,17 @@ export function getAIModel(overrides?: ClientOverrides): ModelConfig {
         }
 
         case "siliconflow": {
-            const apiKey = overrides?.apiKey || process.env.SILICONFLOW_API_KEY
-            const baseURL =
-                overrides?.baseUrl ||
-                process.env.SILICONFLOW_BASE_URL ||
-                "https://api.siliconflow.com/v1"
+            const apiKey = resolveApiKey(overrides, "SILICONFLOW_API_KEY")
+            const serverBaseUrl = resolveBaseUrlEnv(
+                overrides,
+                "SILICONFLOW_BASE_URL",
+            )
+            const baseURL = resolveBaseURL(
+                overrides?.apiKey,
+                overrides?.baseUrl,
+                serverBaseUrl,
+                "https://api.siliconflow.cn/v1",
+            )
             const siliconflowProvider = createOpenAI({
                 apiKey,
                 baseURL,
@@ -724,12 +910,20 @@ export function getAIModel(overrides?: ClientOverrides): ModelConfig {
         }
 
         case "sglang": {
-            const apiKey = overrides?.apiKey || process.env.SGLANG_API_KEY
-            const baseURL = overrides?.baseUrl || process.env.SGLANG_BASE_URL
+            const apiKey = resolveApiKey(overrides, "SGLANG_API_KEY")
+            const serverBaseUrl = resolveBaseUrlEnv(
+                overrides,
+                "SGLANG_BASE_URL",
+            )
+            const baseURL = resolveBaseURL(
+                overrides?.apiKey,
+                overrides?.baseUrl,
+                serverBaseUrl,
+            )
 
             const sglangProvider = createOpenAI({
                 apiKey,
-                baseURL,
+                ...(baseURL && { baseURL }),
                 // Add a custom fetch wrapper to intercept and fix the stream from sglang
                 fetch: async (url, options) => {
                     const response = await fetch(url, options)
@@ -833,9 +1027,16 @@ export function getAIModel(overrides?: ClientOverrides): ModelConfig {
             // Vercel AI Gateway - unified access to multiple AI providers
             // Model format: "provider/model" e.g., "openai/gpt-4o", "anthropic/claude-sonnet-4-5"
             // See: https://vercel.com/ai-gateway
-            const apiKey = overrides?.apiKey || process.env.AI_GATEWAY_API_KEY
-            const baseURL =
-                overrides?.baseUrl || process.env.AI_GATEWAY_BASE_URL
+            const apiKey = resolveApiKey(overrides, "AI_GATEWAY_API_KEY")
+            const serverBaseUrl = resolveBaseUrlEnv(
+                overrides,
+                "AI_GATEWAY_BASE_URL",
+            )
+            const baseURL = resolveBaseURL(
+                overrides?.apiKey,
+                overrides?.baseUrl,
+                serverBaseUrl,
+            )
             // Only use custom configuration if explicitly set (local dev or custom Gateway)
             // Otherwise undefined → AI SDK uses Vercel default (https://ai-gateway.vercel.sh/v1/ai) + OIDC
             if (baseURL || overrides?.apiKey) {
@@ -866,22 +1067,61 @@ export function getAIModel(overrides?: ClientOverrides): ModelConfig {
         }
 
         case "doubao": {
-            const apiKey = overrides?.apiKey || process.env.DOUBAO_API_KEY
-            const baseURL =
-                overrides?.baseUrl ||
-                process.env.DOUBAO_BASE_URL ||
-                "https://ark.cn-beijing.volces.com/api/v3"
-            const doubaoProvider = createDeepSeek({
+            const apiKey = resolveApiKey(overrides, "DOUBAO_API_KEY")
+            const serverBaseUrl = resolveBaseUrlEnv(
+                overrides,
+                "DOUBAO_BASE_URL",
+            )
+            const baseURL = resolveBaseURL(
+                overrides?.apiKey,
+                overrides?.baseUrl,
+                serverBaseUrl,
+                "https://ark.cn-beijing.volces.com/api/v3",
+            )
+            const lowerModelId = modelId.toLowerCase()
+            // Use DeepSeek provider for DeepSeek/Kimi models, OpenAI for others (multimodal support)
+            if (
+                lowerModelId.includes("deepseek") ||
+                lowerModelId.includes("kimi")
+            ) {
+                const doubaoProvider = createDeepSeek({
+                    apiKey,
+                    baseURL,
+                })
+                model = doubaoProvider(modelId)
+            } else {
+                const doubaoProvider = createOpenAI({
+                    apiKey,
+                    baseURL,
+                })
+                model = doubaoProvider.chat(modelId)
+            }
+            break
+        }
+
+        case "modelscope": {
+            const apiKey = resolveApiKey(overrides, "MODELSCOPE_API_KEY")
+            const serverBaseUrl = resolveBaseUrlEnv(
+                overrides,
+                "MODELSCOPE_BASE_URL",
+            )
+            const baseURL = resolveBaseURL(
+                overrides?.apiKey,
+                overrides?.baseUrl,
+                serverBaseUrl,
+                "https://api-inference.modelscope.cn/v1",
+            )
+            const modelscopeProvider = createOpenAI({
                 apiKey,
                 baseURL,
             })
-            model = doubaoProvider(modelId)
+            model = modelscopeProvider.chat(modelId)
             break
         }
 
         default:
             throw new Error(
-                `Unknown AI provider: ${provider}. Supported providers: bedrock, openai, anthropic, google, azure, ollama, openrouter, deepseek, siliconflow, sglang, gateway, edgeone, doubao`,
+                `Unknown AI provider: ${provider}. Supported providers: bedrock, openai, anthropic, google, azure, ollama, openrouter, deepseek, siliconflow, sglang, gateway, edgeone, doubao, modelscope`,
             )
     }
 

lib/i18n/dictionaries/en.json

@@ -28,7 +28,8 @@
         "azure": "Azure OpenAI",
         "openrouter": "OpenRouter",
         "deepseek": "DeepSeek",
-        "siliconflow": "SiliconFlow"
+        "siliconflow": "SiliconFlow",
+        "modelscope": "ModelScope"
     },
     "chat": {
         "placeholder": "Describe your diagram or upload a file...",
@@ -51,7 +52,8 @@
         "badResponse": "Bad response",
         "clickToEdit": "Click to edit",
         "editMessage": "Edit message",
-        "saveAndSubmit": "Save & Submit"
+        "saveAndSubmit": "Save & Submit",
+        "ExtractURL": "Extract from URL"
     },
     "examples": {
         "title": "Create diagrams with AI",
@@ -98,14 +100,22 @@
         "switchTo": "Switch to",
         "minimal": "Minimal",
         "sketch": "Sketch",
-        "closeProtection": "Close Protection",
-        "closeProtectionDescription": "Show confirmation when leaving the page.",
         "diagramStyle": "Diagram Style",
         "diagramStyleDescription": "Toggle between minimal and styled diagram output.",
+        "sendShortcut": "Send Shortcut",
+        "sendShortcutDescription": "Choose how to send messages.",
+        "enterToSend": "Enter to send",
+        "ctrlEnterToSend": "Cmd/Ctrl+Enter to send",
         "diagramActions": "Diagram Actions",
         "diagramActionsDescription": "Manage diagram history and exports",
         "history": "History",
-        "download": "Download"
+        "download": "Download",
+        "proxy": "Proxy Settings",
+        "proxyDescription": "Configure HTTP/HTTPS proxy for API requests (Desktop only)",
+        "httpProxy": "HTTP Proxy",
+        "httpsProxy": "HTTPS Proxy",
+        "applyProxy": "Apply",
+        "proxyApplied": "Proxy settings applied"
     },
     "save": {
         "title": "Save Diagram",
@@ -186,6 +196,15 @@
         "chars": "chars",
         "removeFile": "Remove file"
     },
+    "url": {
+        "title": "Extract Content from URL",
+        "description": "Paste a URL to extract and analyze its content",
+        "Extracting": "Extracting...",
+        "extract": "Extract",
+        "Cancel": "Cancel",
+        "enterUrl": "Please enter a URL",
+        "invalidFormat": "Invalid URL format"
+    },
     "reasoning": {
         "thinking": "Thinking...",
         "thoughtFor": "Thought for {duration} seconds",
@@ -286,10 +305,13 @@
         "noModelsFound": "No models found.",
         "default": "Default",
         "serverDefault": "Server Default",
+        "serverModels": "Server Models",
+        "userModels": "User Models",
         "configureModels": "Configure Models...",
         "onlyVerifiedShown": "Only verified models are shown",
         "showUnvalidatedModels": "Show unvalidated models",
         "allModelsShown": "All models are shown (including unvalidated)",
-        "unvalidatedModelWarning": "This model has not been validated"
+        "unvalidatedModelWarning": "This model has not been validated",
+        "serverDefaultModel": "Server default model"
     }
 }

lib/i18n/dictionaries/ja.json

@@ -28,7 +28,8 @@
         "azure": "Azure OpenAI",
         "openrouter": "OpenRouter",
         "deepseek": "DeepSeek",
-        "siliconflow": "SiliconFlow"
+        "siliconflow": "SiliconFlow",
+        "modelscope": "ModelScope"
     },
     "chat": {
         "placeholder": "ダイアグラムを説明するか、ファイルをアップロード...",
@@ -51,7 +52,8 @@
         "badResponse": "悪い応答",
         "clickToEdit": "クリックして編集",
         "editMessage": "メッセージを編集",
-        "saveAndSubmit": "保存して送信"
+        "saveAndSubmit": "保存して送信",
+        "ExtractURL": "URLから抽出"
     },
     "examples": {
         "title": "AI でダイアグラムを作成",
@@ -98,14 +100,22 @@
         "switchTo": "切り替え",
         "minimal": "ミニマル",
         "sketch": "スケッチ",
-        "closeProtection": "ページ離脱確認",
-        "closeProtectionDescription": "ページを離れる際に確認を表示します。",
         "diagramStyle": "ダイアグラムスタイル",
         "diagramStyleDescription": "ミニマルとスタイル付きの出力を切り替えます。",
+        "sendShortcut": "送信ショートカット",
+        "sendShortcutDescription": "メッセージの送信方法を選択します。",
+        "enterToSend": "Enterで送信",
+        "ctrlEnterToSend": "Cmd/Ctrl+Enterで送信",
         "diagramActions": "ダイアグラム操作",
         "diagramActionsDescription": "ダイアグラムの履歴とエクスポートを管理",
         "history": "履歴",
-        "download": "ダウンロード"
+        "download": "ダウンロード",
+        "proxy": "プロキシ設定",
+        "proxyDescription": "API リクエスト用の HTTP/HTTPS プロキシを設定（デスクトップ版のみ）",
+        "httpProxy": "HTTP プロキシ",
+        "httpsProxy": "HTTPS プロキシ",
+        "applyProxy": "適用",
+        "proxyApplied": "プロキシ設定が適用されました"
     },
     "save": {
         "title": "ダイアグラムを保存",
@@ -186,6 +196,15 @@
         "chars": "文字",
         "removeFile": "ファイルを削除"
     },
+    "url": {
+        "title": "URLからコンテンツを抽出",
+        "description": "URLを貼り付けてそのコンテンツを抽出および分析します",
+        "Extracting": "抽出中...",
+        "extract": "抽出",
+        "Cancel": "キャンセル",
+        "enterUrl": "URLを入力してください",
+        "invalidFormat": "無効なURL形式です"
+    },
     "reasoning": {
         "thinking": "考え中...",
         "thoughtFor": "{duration} 秒考えました",
@@ -286,10 +305,13 @@
         "noModelsFound": "モデルが見つかりません。",
         "default": "デフォルト",
         "serverDefault": "サーバーデフォルト",
+        "serverModels": "サーバーモデル",
+        "userModels": "ユーザーモデル",
         "configureModels": "モデルを設定...",
         "onlyVerifiedShown": "検証済みのモデルのみ表示",
         "showUnvalidatedModels": "未検証のモデルを表示",
         "allModelsShown": "すべてのモデルを表示（未検証を含む）",
-        "unvalidatedModelWarning": "このモデルは検証されていません"
+        "unvalidatedModelWarning": "このモデルは検証されていません",
+        "serverDefaultModel": "サーバーデフォルトモデル"
     }
 }

lib/i18n/dictionaries/zh.json

@@ -28,7 +28,8 @@
         "azure": "Azure OpenAI",
         "openrouter": "OpenRouter",
         "deepseek": "DeepSeek",
-        "siliconflow": "SiliconFlow"
+        "siliconflow": "SiliconFlow",
+        "modelscope": "ModelScope"
     },
     "chat": {
         "placeholder": "描述您的图表或上传文件...",
@@ -51,7 +52,8 @@
         "badResponse": "无帮助",
         "clickToEdit": "点击编辑",
         "editMessage": "编辑消息",
-        "saveAndSubmit": "保存并提交"
+        "saveAndSubmit": "保存并提交",
+        "ExtractURL": "从 URL 提取"
     },
     "examples": {
         "title": "用 AI 创建图表",
@@ -98,14 +100,22 @@
         "switchTo": "切换到",
         "minimal": "简约",
         "sketch": "草图",
-        "closeProtection": "关闭确认",
-        "closeProtectionDescription": "离开页面时显示确认。",
         "diagramStyle": "图表样式",
         "diagramStyleDescription": "切换简约与精致图表输出模式。",
+        "sendShortcut": "发送快捷键",
+        "sendShortcutDescription": "选择发送消息的方式。",
+        "enterToSend": "回车发送",
+        "ctrlEnterToSend": "Cmd/Ctrl+回车发送",
         "diagramActions": "图表操作",
         "diagramActionsDescription": "管理图表历史记录和导出",
         "history": "历史记录",
-        "download": "下载"
+        "download": "下载",
+        "proxy": "代理设置",
+        "proxyDescription": "配置 API 请求的 HTTP/HTTPS 代理（仅桌面版）",
+        "httpProxy": "HTTP 代理",
+        "httpsProxy": "HTTPS 代理",
+        "applyProxy": "应用",
+        "proxyApplied": "代理设置已应用"
     },
     "save": {
         "title": "保存图表",
@@ -186,6 +196,15 @@
         "chars": "字符",
         "removeFile": "移除文件"
     },
+    "url": {
+        "title": "从 URL 提取内容",
+        "description": "粘贴 URL 以提取和分析其内容",
+        "Extracting": "提取中...",
+        "extract": "提取",
+        "Cancel": "取消",
+        "enterUrl": "请输入 URL",
+        "invalidFormat": "URL 格式无效"
+    },
     "reasoning": {
         "thinking": "思考中...",
         "thoughtFor": "思考了 {duration} 秒",
@@ -286,10 +305,13 @@
         "noModelsFound": "未找到模型。",
         "default": "默认",
         "serverDefault": "服务器默认",
+        "serverModels": "服务器模型",
+        "userModels": "用户模型",
         "configureModels": "配置模型...",
         "onlyVerifiedShown": "仅显示已验证的模型",
         "showUnvalidatedModels": "显示未验证的模型",
         "allModelsShown": "显示所有模型（包括未验证的）",
-        "unvalidatedModelWarning": "此模型尚未验证"
+        "unvalidatedModelWarning": "此模型尚未验证",
+        "serverDefaultModel": "服务器默认模型"
     }
 }

lib/server-model-config.ts

@@ -0,0 +1,151 @@
+import fs from "fs/promises"
+import path from "path"
+import { z } from "zod"
+import type { ProviderName } from "@/lib/types/model-config"
+import { PROVIDER_INFO } from "@/lib/types/model-config"
+
+export const ProviderNameSchema: z.ZodType<ProviderName> = z
+    .string()
+    .refine((val): val is ProviderName => val in PROVIDER_INFO, {
+        message: "Invalid provider name",
+    })
+
+export const ServerProviderSchema = z.object({
+    name: z.string().min(1),
+    provider: ProviderNameSchema,
+    models: z.array(z.string().min(1)),
+    // Optional: custom environment variable name for API key
+    // e.g., "OPENAI_API_KEY_TEAM_A" instead of default "OPENAI_API_KEY"
+    apiKeyEnv: z.string().min(1).optional(),
+    // Optional: custom environment variable name for base URL
+    baseUrlEnv: z.string().min(1).optional(),
+    // Optional: mark the first model in this provider as the default
+    default: z.boolean().optional(),
+})
+
+export const ServerModelsConfigSchema = z.object({
+    providers: z.array(ServerProviderSchema),
+})
+
+export type ServerProviderConfig = z.infer<typeof ServerProviderSchema>
+export type ServerModelsConfig = z.infer<typeof ServerModelsConfigSchema>
+
+export interface FlattenedServerModel {
+    id: string // "server:<slugified-name>:<modelId>" - name ensures uniqueness for multiple API keys per provider
+    modelId: string
+    provider: ProviderName
+    providerLabel: string
+    isDefault: boolean
+    // Custom env var names for credentials (optional)
+    apiKeyEnv?: string
+    baseUrlEnv?: string
+}
+
+/**
+ * Convert provider name to URL-safe slug for use in model ID
+ * e.g., "OpenAI Production" → "openai-production"
+ */
+function slugify(name: string): string {
+    return name
+        .toLowerCase()
+        .replace(/[^a-z0-9]+/g, "-")
+        .replace(/^-|-$/g, "")
+}
+
+function getConfigPath(): string {
+    const custom = process.env.AI_MODELS_CONFIG_PATH
+    if (custom && custom.trim().length > 0) return custom
+    return path.join(process.cwd(), "ai-models.json")
+}
+
+export async function loadRawServerModelsConfig(): Promise<ServerModelsConfig | null> {
+    // Priority 1: AI_MODELS_CONFIG env var (JSON string) - for cloud deployments
+    const envConfig = process.env.AI_MODELS_CONFIG
+    if (envConfig && envConfig.trim().length > 0) {
+        try {
+            const json = JSON.parse(envConfig)
+            return ServerModelsConfigSchema.parse(json)
+        } catch (err) {
+            console.error(
+                "[server-model-config] Failed to parse AI_MODELS_CONFIG:",
+                err,
+            )
+            return null
+        }
+    }
+
+    // Priority 2: ai-models.json file
+    const configPath = getConfigPath()
+    try {
+        const jsonStr = await fs.readFile(configPath, "utf8")
+        const json = JSON.parse(jsonStr)
+        return ServerModelsConfigSchema.parse(json)
+    } catch (err: any) {
+        if (err?.code === "ENOENT") {
+            return null
+        }
+        console.error(
+            "[server-model-config] Failed to load ai-models.json:",
+            err,
+        )
+        return null
+    }
+}
+
+export async function loadFlattenedServerModels(): Promise<
+    FlattenedServerModel[]
+> {
+    const cfg = await loadRawServerModelsConfig()
+    if (!cfg) return []
+
+    const defaultProvider = process.env.AI_PROVIDER as ProviderName | undefined
+    const defaultModelId = process.env.AI_MODEL
+
+    const flattened: FlattenedServerModel[] = []
+
+    for (const p of cfg.providers) {
+        const providerLabel =
+            p.name || PROVIDER_INFO[p.provider]?.label || p.provider
+
+        // Use slugified name for unique ID (supports multiple API keys per provider)
+        const nameSlug = slugify(p.name)
+
+        for (const modelId of p.models) {
+            const id = `server:${nameSlug}:${modelId}`
+
+            // Default model priority:
+            // 1. From ai-models.json: first model of provider with default: true
+            // 2. From env vars: AI_MODEL matches (legacy behavior)
+            const isDefault =
+                (p.default === true && modelId === p.models[0]) ||
+                (!!defaultModelId &&
+                    modelId === defaultModelId &&
+                    (!defaultProvider || defaultProvider === p.provider))
+
+            flattened.push({
+                id,
+                modelId,
+                provider: p.provider,
+                providerLabel,
+                isDefault,
+                apiKeyEnv: p.apiKeyEnv,
+                baseUrlEnv: p.baseUrlEnv,
+            })
+        }
+    }
+
+    return flattened
+}
+
+/**
+ * Find a server model by its ID (format: "server:<slugified-name>:<modelId>")
+ * Returns the model config including apiKeyEnv/baseUrlEnv if configured
+ */
+export async function findServerModelById(
+    modelId: string,
+): Promise<FlattenedServerModel | null> {
+    if (!modelId.startsWith("server:")) return null
+
+    const models = await loadFlattenedServerModels()
+    return models.find((m) => m.id === modelId) || null
+}

lib/storage.ts

@@ -12,7 +12,6 @@ export const STORAGE_KEYS = {
 
     // Settings
     accessCode: "next-ai-draw-io-access-code",
-    closeProtection: "next-ai-draw-io-close-protection",
     accessCodeRequired: "next-ai-draw-io-access-code-required",
     aiProvider: "next-ai-draw-io-ai-provider",
     aiBaseUrl: "next-ai-draw-io-ai-base-url",
@@ -22,4 +21,7 @@ export const STORAGE_KEYS = {
     // Multi-model configuration
     modelConfigs: "next-ai-draw-io-model-configs",
     selectedModelId: "next-ai-draw-io-selected-model-id",
+
+    // Chat input preferences
+    sendShortcut: "next-ai-draw-io-send-shortcut",
 } as const

lib/token-counter.ts

@@ -1,39 +0,0 @@
-/**
- * Token counting utilities using js-tiktoken
- *
- * Uses cl100k_base encoding (GPT-4) which is close to Claude's tokenization.
- * This is a pure JavaScript implementation, no WASM required.
- */
-
-import { encodingForModel } from "js-tiktoken"
-import { DEFAULT_SYSTEM_PROMPT, EXTENDED_SYSTEM_PROMPT } from "./system-prompts"
-
-const encoder = encodingForModel("gpt-4o")
-
-/**
- * Count the number of tokens in a text string
- * @param text - The text to count tokens for
- * @returns The number of tokens
- */
-export function countTextTokens(text: string): number {
-    return encoder.encode(text).length
-}
-
-/**
- * Get token counts for the system prompts
- * Useful for debugging and optimizing prompt sizes
- * @returns Object with token counts for default and extended prompts
- */
-export function getSystemPromptTokenCounts(): {
-    default: number
-    extended: number
-    additions: number
-} {
-    const defaultTokens = countTextTokens(DEFAULT_SYSTEM_PROMPT)
-    const extendedTokens = countTextTokens(EXTENDED_SYSTEM_PROMPT)
-    return {
-        default: defaultTokens,
-        extended: extendedTokens,
-        additions: extendedTokens - defaultTokens,
-    }
-}

lib/types/model-config.ts

@@ -4,8 +4,10 @@ export type ProviderName =
     | "openai"
     | "anthropic"
     | "google"
+    | "vertexai"
     | "azure"
     | "bedrock"
+    | "ollama"
     | "openrouter"
     | "deepseek"
     | "siliconflow"
@@ -13,6 +15,7 @@ export type ProviderName =
     | "gateway"
     | "edgeone"
     | "doubao"
+    | "modelscope"
 
 // Individual model configuration
 export interface ModelConfig {
@@ -34,6 +37,9 @@ export interface ProviderConfig {
     awsSecretAccessKey?: string
     awsRegion?: string
     awsSessionToken?: string // Optional, for temporary credentials
+    // Vertex AI specific fields
+    vertexApiKey?: string // Express Mode API key
+
     models: ModelConfig[]
     validated?: boolean // Has API key been validated
 }
@@ -48,7 +54,7 @@ export interface MultiModelConfig {
 
 // Flattened model for dropdown display
 export interface FlattenedModel {
-    id: string // Model config UUID
+    id: string // Model config UUID or synthetic server ID (e.g., "server:provider:modelId")
     modelId: string // Actual model ID
     provider: ProviderName
     providerLabel: string // Provider display name
@@ -59,7 +65,17 @@ export interface FlattenedModel {
     awsSecretAccessKey?: string
     awsRegion?: string
     awsSessionToken?: string
+    // Vertex AI specific fields
+    vertexApiKey?: string // Express Mode API key
+
     validated?: boolean // Has this model been validated
+    // Source of this model config: user-defined (client) or server-defined
+    source?: "user" | "server"
+    // Whether this model is the server default (matches AI_MODEL env var)
+    isDefault?: boolean
+    // Custom env var names for server models (allows multiple API keys per provider)
+    apiKeyEnv?: string
+    baseUrlEnv?: string
 }
 
 // Provider metadata
@@ -73,13 +89,18 @@ export const PROVIDER_INFO: Record<
         defaultBaseUrl: "https://api.anthropic.com/v1",
     },
     google: { label: "Google" },
+    vertexai: { label: "Google Vertex AI" },
     azure: { label: "Azure OpenAI" },
     bedrock: { label: "Amazon Bedrock" },
+    ollama: {
+        label: "Ollama",
+        defaultBaseUrl: "http://localhost:11434",
+    },
     openrouter: { label: "OpenRouter" },
     deepseek: { label: "DeepSeek" },
     siliconflow: {
         label: "SiliconFlow",
-        defaultBaseUrl: "https://api.siliconflow.com/v1",
+        defaultBaseUrl: "https://api.siliconflow.cn/v1",
     },
     sglang: {
         label: "SGLang",
@@ -91,10 +112,14 @@ export const PROVIDER_INFO: Record<
         label: "Doubao (ByteDance)",
         defaultBaseUrl: "https://ark.cn-beijing.volces.com/api/v3",
     },
+    modelscope: {
+        label: "ModelScope",
+        defaultBaseUrl: "https://api-inference.modelscope.cn/v1",
+    },
 }
 
 // Suggested models per provider for quick add
-export const SUGGESTED_MODELS: Record<ProviderName, string[]> = {
+export const SUGGESTED_MODELS: Partial<Record<ProviderName, string[]>> = {
     openai: [
         "gpt-5.2-pro",
         "gpt-5.2-chat-latest",
@@ -147,6 +172,17 @@ export const SUGGESTED_MODELS: Record<ProviderName, string[]> = {
         // Legacy
         "gemini-pro",
     ],
+    vertexai: [
+        // Gemini 2.5 series
+        "gemini-2.5-pro",
+        "gemini-2.5-flash",
+        // Gemini 2.0 series
+        "gemini-2.0-flash",
+        "gemini-2.0-flash-exp",
+        // Gemini 1.5 series
+        "gemini-1.5-pro",
+        "gemini-1.5-flash",
+    ],
     azure: ["gpt-4o", "gpt-4o-mini", "gpt-4-turbo", "gpt-4", "gpt-35-turbo"],
     bedrock: [
         // Anthropic Claude
@@ -231,6 +267,17 @@ export const SUGGESTED_MODELS: Record<ProviderName, string[]> = {
         "doubao-pro-32k-241215",
         "doubao-pro-256k-241215",
     ],
+    modelscope: [
+        // Qwen
+        "Qwen/Qwen2.5-72B-Instruct",
+        "Qwen/Qwen2.5-32B-Instruct",
+        "Qwen/Qwen3-235B-A22B-Instruct-2507",
+        "Qwen/Qwen3-VL-235B-A22B-Instruct",
+        "Qwen/Qwen3-32B",
+        // DeepSeek
+        "deepseek-ai/DeepSeek-R1-0528",
+        "deepseek-ai/DeepSeek-V3.2",
+    ],
 }
 
 // Helper to generate UUID
@@ -267,7 +314,7 @@ export function createModelConfig(modelId: string): ModelConfig {
     }
 }
 
-// Get all models as flattened list for dropdown
+// Get all models as flattened list for dropdown (user-defined only)
 export function flattenModels(config: MultiModelConfig): FlattenedModel[] {
     const models: FlattenedModel[] = []
 
@@ -289,7 +336,12 @@ export function flattenModels(config: MultiModelConfig): FlattenedModel[] {
                 awsSecretAccessKey: provider.awsSecretAccessKey,
                 awsRegion: provider.awsRegion,
                 awsSessionToken: provider.awsSessionToken,
+                // Vertex AI fields
+                vertexApiKey: provider.vertexApiKey,
+
                 validated: model.validated,
+                source: "user",
+                isDefault: false,
             })
         }
     }

lib/url-utils.ts

@@ -0,0 +1,49 @@
+import { z } from "zod"
+
+export interface UrlData {
+    url: string
+    title: string
+    content: string
+    charCount: number
+    isExtracting: boolean
+}
+
+const UrlResponseSchema = z.object({
+    title: z.string().default("Untitled"),
+    content: z.string(),
+    charCount: z.number().int().nonnegative(),
+})
+
+export async function extractUrlContent(url: string): Promise<UrlData> {
+    const response = await fetch("/api/parse-url", {
+        method: "POST",
+        headers: { "Content-Type": "application/json" },
+        body: JSON.stringify({ url }),
+    })
+
+    // Try to parse JSON once
+    const raw = await response
+        .json()
+        .catch(() => ({ error: "Unexpected non-JSON response" }))
+
+    if (!response.ok) {
+        const message =
+            typeof raw === "object" && raw && "error" in raw
+                ? String((raw as any).error)
+                : "Failed to extract URL content"
+        throw new Error(message)
+    }
+
+    const parsed = UrlResponseSchema.safeParse(raw)
+    if (!parsed.success) {
+        throw new Error("Malformed response from URL extraction API")
+    }
+
+    return {
+        url,
+        title: parsed.data.title,
+        content: parsed.data.content,
+        charCount: parsed.data.charCount,
+        isExtracting: false,
+    }
+}

lib/utils.ts

@@ -1,6 +1,9 @@
 import { type ClassValue, clsx } from "clsx"
 import * as pako from "pako"
 import { twMerge } from "tailwind-merge"
+import type { DiagramOperation } from "@/components/chat/types"
+
+export type { DiagramOperation }
 
 export function cn(...inputs: ClassValue[]) {
     return twMerge(clsx(inputs))
@@ -473,12 +476,6 @@ export function replaceNodes(currentXML: string, nodes: string): string {
 // ID-based Diagram Operations
 // ============================================================================
 
-export interface DiagramOperation {
-    operation: "update" | "add" | "delete"
-    cell_id: string
-    new_xml?: string
-}
-
 export interface OperationError {
     type: "update" | "add" | "delete"
     cellId: string

package.json

@@ -1,6 +1,6 @@
 {
     "name": "next-ai-draw-io",
-    "version": "0.4.9",
+    "version": "0.4.10",
     "license": "Apache-2.0",
     "private": true,
     "main": "dist-electron/main/index.js",
@@ -24,6 +24,7 @@
         "dist": "npm run electron:build && npm run electron:prepare && npx electron-builder --config electron/electron-builder.yml",
         "dist:mac": "npm run electron:build && npm run electron:prepare && npx electron-builder --config electron/electron-builder.yml --mac",
         "dist:win": "npm run electron:build && npm run electron:prepare && npx electron-builder --config electron/electron-builder.yml --win",
+        "dist:win:build": "npm run electron:build && npm run electron:prepare && npx electron-builder --config electron/electron-builder.yml --win --publish never",
         "dist:linux": "npm run electron:build && npm run electron:prepare && npx electron-builder --config electron/electron-builder.yml --linux",
         "dist:all": "npm run electron:build && npm run electron:prepare && npx electron-builder --config electron/electron-builder.yml --mac --win --linux",
         "test": "vitest",
@@ -36,18 +37,21 @@
         "@ai-sdk/deepseek": "^2.0.0",
         "@ai-sdk/gateway": "^3.0.0",
         "@ai-sdk/google": "^3.0.0",
+        "@ai-sdk/google-vertex": "^4.0.16",
         "@ai-sdk/openai": "^3.0.0",
         "@ai-sdk/react": "^3.0.1",
         "@aws-sdk/client-dynamodb": "^3.957.0",
         "@aws-sdk/credential-providers": "^3.943.0",
+        "@extractus/article-extractor": "^8.0.18",
         "@formatjs/intl-localematcher": "^0.7.2",
         "@langfuse/client": "^4.4.9",
         "@langfuse/otel": "^4.4.4",
         "@langfuse/tracing": "^4.4.9",
         "@next/third-parties": "^16.0.6",
-        "@opennextjs/cloudflare": "1.14.7",
+        "@opennextjs/cloudflare": "1.14.8",
         "@openrouter/ai-sdk-provider": "^1.5.4",
-        "@opentelemetry/exporter-trace-otlp-http": "^0.208.0",
+        "@opentelemetry/api": "^1.9.0",
+        "@opentelemetry/exporter-trace-otlp-http": "^0.209.0",
         "@opentelemetry/sdk-trace-node": "^2.2.0",
         "@radix-ui/react-alert-dialog": "^1.1.15",
         "@radix-ui/react-collapsible": "^1.1.12",
@@ -67,10 +71,10 @@
         "clsx": "^2.1.1",
         "cmdk": "^1.1.1",
         "idb": "^8.0.3",
-        "js-tiktoken": "^1.0.21",
         "jsonrepair": "^3.13.1",
         "lucide-react": "^0.562.0",
         "motion": "^12.23.25",
+        "nanoid": "^5.0.0",
         "negotiator": "^1.0.0",
         "next": "^16.0.7",
         "ollama-ai-provider-v2": "^2.0.0",
@@ -87,6 +91,7 @@
         "sonner": "^2.0.7",
         "tailwind-merge": "^3.0.2",
         "tailwindcss-animate": "^1.0.7",
+        "turndown": "^7.2.0",
         "unpdf": "^1.4.0",
         "zod": "^4.1.12"
     },
@@ -115,6 +120,7 @@
         "@types/pako": "^2.0.3",
         "@types/react": "^19",
         "@types/react-dom": "^19",
+        "@types/turndown": "^5.0.6",
         "@vitejs/plugin-react": "^5.1.2",
         "@vitest/coverage-v8": "^4.0.16",
         "concurrently": "^9.2.1",
@@ -133,7 +139,7 @@
         "vite-tsconfig-paths": "^6.0.3",
         "vitest": "^4.0.16",
         "wait-on": "^9.0.3",
-        "wrangler": "4.54.0"
+        "wrangler": "4.58.0"
     },
     "overrides": {
         "@openrouter/ai-sdk-provider": {

packages/claude-plugin/.claude-plugin/plugin.json

@@ -0,0 +1,11 @@
+{
+    "name": "next-ai-drawio",
+    "version": "1.0.0",
+    "description": "AI-powered Draw.io diagram generation with real-time browser preview. Create flowcharts, architecture diagrams, and more through natural language.",
+    "author": {
+        "name": "DayuanJiang"
+    },
+    "repository": "https://github.com/DayuanJiang/next-ai-draw-io",
+    "homepage": "https://next-ai-drawio.jiang.jp",
+    "license": "Apache-2.0"
+}

packages/claude-plugin/.mcp.json

@@ -0,0 +1,8 @@
+{
+    "mcpServers": {
+        "drawio": {
+            "command": "npx",
+            "args": ["@next-ai-drawio/mcp-server@latest"]
+        }
+    }
+}

packages/claude-plugin/README.md

@@ -0,0 +1,107 @@
+# Next AI Draw.io - Claude Code Plugin
+
+AI-powered Draw.io diagram generation with real-time browser preview for Claude Code.
+
+## Installation
+
+### From Plugin Directory (Coming Soon)
+
+Once approved, install via:
+```
+/plugin install next-ai-drawio
+```
+
+### Manual Installation
+
+```bash
+claude --plugin-dir /path/to/packages/claude-plugin
+```
+
+Or add the MCP server directly:
+```bash
+claude mcp add drawio -- npx @next-ai-drawio/mcp-server@latest
+```
+
+## Features
+
+- **Real-time Preview**: Diagrams appear and update in your browser as Claude creates them
+- **Version History**: Restore previous diagram versions with visual thumbnails
+- **Natural Language**: Describe diagrams in plain text - flowcharts, architecture diagrams, etc.
+- **Edit Support**: Modify existing diagrams with natural language instructions
+- **Export**: Save diagrams as `.drawio` files
+- **Self-contained**: Embedded server, no external dependencies required
+
+## Use Case Examples
+
+### 1. Create Architecture Diagrams
+
+```
+Generate an AWS architecture diagram with Lambda, API Gateway, DynamoDB,
+and S3 for a serverless REST API
+```
+
+### 2. Flowchart Generation
+
+```
+Create a flowchart showing the CI/CD pipeline: code commit -> build ->
+test -> staging deploy -> production deploy with approval gates
+```
+
+### 3. System Design Documentation
+
+```
+Design a microservices e-commerce system with user service, product catalog,
+shopping cart, order processing, and payment gateway
+```
+
+### 4. Cloud Architecture (AWS/GCP/Azure)
+
+```
+Generate a GCP architecture diagram with Cloud Run, Cloud SQL, and
+Cloud Storage for a web application
+```
+
+### 5. Sequence Diagrams
+
+```
+Create a sequence diagram showing OAuth 2.0 authorization code flow
+between user, client app, auth server, and resource server
+```
+
+## Available Tools
+
+| Tool | Description |
+|------|-------------|
+| `start_session` | Opens browser with real-time diagram preview |
+| `create_new_diagram` | Create a new diagram from XML |
+| `edit_diagram` | Edit diagram by ID-based operations |
+| `get_diagram` | Get the current diagram XML |
+| `export_diagram` | Save diagram to a `.drawio` file |
+
+## How It Works
+
+```
+Claude Code <--stdio--> MCP Server <--http--> Browser (draw.io)
+```
+
+1. Ask Claude to create a diagram
+2. Claude calls `start_session` to open a browser window
+3. Claude generates diagram XML and sends it to the browser
+4. You see the diagram update in real-time!
+
+## Configuration
+
+| Variable | Default | Description |
+|----------|---------|-------------|
+| `PORT` | `6002` | Port for the embedded HTTP server |
+| `DRAWIO_BASE_URL` | `https://embed.diagrams.net` | Base URL for draw.io (for self-hosted deployments) |
+
+## Links
+
+- [Homepage](https://next-ai-drawio.jiang.jp)
+- [GitHub Repository](https://github.com/DayuanJiang/next-ai-draw-io)
+- [MCP Server Documentation](https://github.com/DayuanJiang/next-ai-draw-io/tree/main/packages/mcp-server)
+
+## License
+
+Apache-2.0

packages/mcp-server/package-lock.json

@@ -1,18 +1,18 @@
 {
     "name": "@next-ai-drawio/mcp-server",
-    "version": "0.1.6",
+    "version": "0.1.12",
     "lockfileVersion": 3,
     "requires": true,
     "packages": {
         "": {
             "name": "@next-ai-drawio/mcp-server",
-            "version": "0.1.6",
+            "version": "0.1.12",
             "license": "Apache-2.0",
             "dependencies": {
                 "@modelcontextprotocol/sdk": "^1.0.4",
                 "linkedom": "^0.18.0",
                 "open": "^11.0.0",
-                "zod": "^3.24.0"
+                "zod": "^4.0.0"
             },
             "bin": {
                 "next-ai-drawio-mcp": "dist/index.js"
@@ -481,9 +481,9 @@
             }
         },
         "node_modules/@modelcontextprotocol/sdk": {
-            "version": "1.25.1",
-            "resolved": "https://registry.npmjs.org/@modelcontextprotocol/sdk/-/sdk-1.25.1.tgz",
-            "integrity": "sha512-yO28oVFFC7EBoiKdAn+VqRm+plcfv4v0xp6osG/VsCB0NlPZWi87ajbCZZ8f/RvOFLEu7//rSRmuZZ7lMoe3gQ==",
+            "version": "1.25.2",
+            "resolved": "https://registry.npmjs.org/@modelcontextprotocol/sdk/-/sdk-1.25.2.tgz",
+            "integrity": "sha512-LZFeo4F9M5qOhC/Uc1aQSrBHxMrvxett+9KLHt7OhcExtoiRN9DKgbZffMP/nxjutWDQpfMDfP3nkHI4X9ijww==",
             "license": "MIT",
             "dependencies": {
                 "@hono/node-server": "^1.19.7",
@@ -520,9 +520,9 @@
             }
         },
         "node_modules/@types/node": {
-            "version": "24.10.4",
-            "resolved": "https://registry.npmjs.org/@types/node/-/node-24.10.4.tgz",
-            "integrity": "sha512-vnDVpYPMzs4wunl27jHrfmwojOGKya0xyM3sH+UE5iv5uPS6vX7UIoh6m+vQc5LGBq52HBKPIn/zcSZVzeDEZg==",
+            "version": "24.10.6",
+            "resolved": "https://registry.npmjs.org/@types/node/-/node-24.10.6.tgz",
+            "integrity": "sha512-B8h60xgJMR/xmgyX9fncRzEW9gCxoJjdenUhke2v1JGOd/V66KopmWrLPXi5oUI4VuiGK+d+HlXJjDRZMj21EQ==",
             "dev": true,
             "license": "MIT",
             "dependencies": {
@@ -2051,9 +2051,9 @@
             }
         },
         "node_modules/zod": {
-            "version": "3.25.76",
-            "resolved": "https://registry.npmjs.org/zod/-/zod-3.25.76.tgz",
-            "integrity": "sha512-gzUt/qt81nXsFGKIFcC3YnfEAx5NkunCfnDlvuBSSFS02bcXu4Lmea0AFIUwbLWxWPx3d9p8S5QoaujKcNQxcQ==",
+            "version": "4.3.5",
+            "resolved": "https://registry.npmjs.org/zod/-/zod-4.3.5.tgz",
+            "integrity": "sha512-k7Nwx6vuWx1IJ9Bjuf4Zt1PEllcwe7cls3VNzm4CQ1/hgtFUK2bRNG3rvnpPUhFjmqJKAKtjV576KnUkHocg/g==",
             "license": "MIT",
             "peer": true,
             "funding": {

packages/mcp-server/package.json

@@ -1,6 +1,6 @@
 {
     "name": "@next-ai-drawio/mcp-server",
-    "version": "0.1.11",
+    "version": "0.1.13",
     "description": "MCP server for Next AI Draw.io - AI-powered diagram generation with real-time browser preview",
     "type": "module",
     "main": "dist/index.js",
@@ -21,16 +21,16 @@
         "claude",
         "model-context-protocol"
     ],
-    "author": "Biki-dev",
+    "author": "DayuanJiang",
     "license": "Apache-2.0",
     "repository": {
         "type": "git",
-        "url": "https://github.com/Biki-dev/next-ai-draw-io",
+        "url": "https://github.com/DayuanJiang/next-ai-draw-io",
         "directory": "packages/mcp-server"
     },
     "homepage": "https://next-ai-drawio.jiang.jp",
     "bugs": {
-        "url": "https://github.com/Biki-dev/next-ai-draw-io/issues"
+        "url": "https://github.com/DayuanJiang/next-ai-draw-io/issues"
     },
     "publishConfig": {
         "access": "public"
@@ -39,7 +39,7 @@
         "@modelcontextprotocol/sdk": "^1.0.4",
         "linkedom": "^0.18.0",
         "open": "^11.0.0",
-        "zod": "^3.24.0"
+        "zod": "^4.0.0"
     },
     "devDependencies": {
         "@types/node": "^24.0.0",

packages/mcp-server/src/index.ts

@@ -260,6 +260,7 @@ COMMON STYLES:
             // Update session state
             currentSession.xml = xml
             currentSession.version++
+            currentSession.lastGetDiagramTime = Date.now()
 
             // Push to embedded server state
             setState(currentSession.id, xml)

scripts/prepare-electron-build.mjs

@@ -38,4 +38,12 @@ const targetStaticDir = join(targetDir, ".next", "static")
 mkdirSync(targetStaticDir, { recursive: true })
 cpSync(staticDir, targetStaticDir, { recursive: true })
 
+// Copy public folder (required for favicon-white.svg and other assets)
+console.log("Copying public folder...")
+const publicDir = join(rootDir, "public")
+const targetPublicDir = join(targetDir, "public")
+if (existsSync(publicDir)) {
+    cpSync(publicDir, targetPublicDir, { recursive: true })
+}
+
 console.log("Done! Files prepared in electron-standalone/")

tests/unit/ai-providers.test.ts

@@ -1,5 +1,141 @@
 import { describe, expect, it } from "vitest"
-import { supportsImageInput, supportsPromptCaching } from "@/lib/ai-providers"
+import {
+    resolveBaseURL,
+    supportsImageInput,
+    supportsPromptCaching,
+} from "@/lib/ai-providers"
+
+describe("resolveBaseURL", () => {
+    const SERVER_BASE_URL = "https://server-proxy.example.com"
+    const USER_BASE_URL = "https://user-proxy.example.com"
+    const DEFAULT_BASE_URL = "https://api.provider.com/v1"
+    const USER_API_KEY = "user-api-key-123"
+
+    describe("when user provides their own API key", () => {
+        it("uses user's baseUrl when provided", () => {
+            const result = resolveBaseURL(
+                USER_API_KEY,
+                USER_BASE_URL,
+                SERVER_BASE_URL,
+                DEFAULT_BASE_URL,
+            )
+            expect(result).toBe(USER_BASE_URL)
+        })
+
+        it("uses default baseUrl when user provides no baseUrl", () => {
+            const result = resolveBaseURL(
+                USER_API_KEY,
+                null,
+                SERVER_BASE_URL,
+                DEFAULT_BASE_URL,
+            )
+            expect(result).toBe(DEFAULT_BASE_URL)
+        })
+
+        it("returns undefined when user provides no baseUrl and no default exists", () => {
+            const result = resolveBaseURL(
+                USER_API_KEY,
+                null,
+                SERVER_BASE_URL,
+                undefined,
+            )
+            expect(result).toBeUndefined()
+        })
+
+        it("does NOT use server's baseUrl even when available", () => {
+            const result = resolveBaseURL(
+                USER_API_KEY,
+                undefined,
+                SERVER_BASE_URL,
+                undefined,
+            )
+            // Should NOT return SERVER_BASE_URL
+            expect(result).not.toBe(SERVER_BASE_URL)
+            expect(result).toBeUndefined()
+        })
+
+        it("prefers user's baseUrl over default", () => {
+            const result = resolveBaseURL(
+                USER_API_KEY,
+                USER_BASE_URL,
+                SERVER_BASE_URL,
+                DEFAULT_BASE_URL,
+            )
+            expect(result).toBe(USER_BASE_URL)
+        })
+    })
+
+    describe("when using server credentials (no user API key)", () => {
+        it("uses user's baseUrl when provided (overrides server)", () => {
+            const result = resolveBaseURL(
+                null,
+                USER_BASE_URL,
+                SERVER_BASE_URL,
+                DEFAULT_BASE_URL,
+            )
+            expect(result).toBe(USER_BASE_URL)
+        })
+
+        it("falls back to server's baseUrl when no user baseUrl", () => {
+            const result = resolveBaseURL(
+                null,
+                null,
+                SERVER_BASE_URL,
+                DEFAULT_BASE_URL,
+            )
+            expect(result).toBe(SERVER_BASE_URL)
+        })
+
+        it("falls back to default when no user or server baseUrl", () => {
+            const result = resolveBaseURL(
+                null,
+                null,
+                undefined,
+                DEFAULT_BASE_URL,
+            )
+            expect(result).toBe(DEFAULT_BASE_URL)
+        })
+
+        it("returns undefined when no baseUrl available anywhere", () => {
+            const result = resolveBaseURL(null, null, undefined, undefined)
+            expect(result).toBeUndefined()
+        })
+
+        it("handles undefined apiKey same as null", () => {
+            const result = resolveBaseURL(
+                undefined,
+                null,
+                SERVER_BASE_URL,
+                DEFAULT_BASE_URL,
+            )
+            expect(result).toBe(SERVER_BASE_URL)
+        })
+    })
+
+    describe("edge cases", () => {
+        it("handles empty string apiKey as falsy (uses server config)", () => {
+            const result = resolveBaseURL(
+                "",
+                null,
+                SERVER_BASE_URL,
+                DEFAULT_BASE_URL,
+            )
+            // Empty string is falsy, so should use server config
+            expect(result).toBe(SERVER_BASE_URL)
+        })
+
+        it("handles empty string baseUrl as falsy", () => {
+            const result = resolveBaseURL(
+                USER_API_KEY,
+                "",
+                SERVER_BASE_URL,
+                DEFAULT_BASE_URL,
+            )
+            // Empty string baseUrl is falsy, should fall back to default
+            expect(result).toBe(DEFAULT_BASE_URL)
+        })
+    })
+})
 
 describe("supportsPromptCaching", () => {
     it("returns true for Claude models", () => {

tests/unit/server-model-config.test.ts

@@ -0,0 +1,85 @@
+import { afterEach, describe, expect, it } from "vitest"
+import {
+    loadFlattenedServerModels,
+    type ServerModelsConfig,
+    ServerModelsConfigSchema,
+} from "@/lib/server-model-config"
+
+const ORIGINAL_ENV = { ...process.env }
+
+afterEach(() => {
+    process.env.AI_PROVIDER = ORIGINAL_ENV.AI_PROVIDER
+    process.env.AI_MODEL = ORIGINAL_ENV.AI_MODEL
+    process.env.AI_MODELS_CONFIG_PATH = ORIGINAL_ENV.AI_MODELS_CONFIG_PATH
+    process.env.AI_MODELS_CONFIG = ORIGINAL_ENV.AI_MODELS_CONFIG
+})
+
+describe("ServerModelsConfigSchema", () => {
+    it("accepts valid provider names", () => {
+        const config: ServerModelsConfig = {
+            providers: [
+                {
+                    name: "OpenAI Server",
+                    provider: "openai",
+                    models: ["gpt-4o"],
+                },
+            ],
+        }
+
+        expect(() => ServerModelsConfigSchema.parse(config)).not.toThrow()
+    })
+
+    it("rejects invalid provider names", () => {
+        const invalidConfig = {
+            providers: [
+                {
+                    name: "Invalid Provider",
+                    // Cast to any so we can verify runtime validation, not TypeScript
+                    provider: "invalid-provider" as any,
+                    models: ["model-1"],
+                },
+            ],
+        }
+
+        expect(() =>
+            ServerModelsConfigSchema.parse(invalidConfig as any),
+        ).toThrow()
+    })
+})
+
+describe("loadFlattenedServerModels", () => {
+    it("returns empty array when config file is missing", async () => {
+        // Point to a non-existent config path so fs.readFile throws ENOENT
+        process.env.AI_MODELS_CONFIG_PATH = `non-existent-config-${Date.now()}.json`
+
+        const models = await loadFlattenedServerModels()
+        expect(models).toEqual([])
+    })
+
+    it("flattens providers and marks default model from env var config", async () => {
+        // Use AI_MODELS_CONFIG env var instead of file
+        const config: ServerModelsConfig = {
+            providers: [
+                {
+                    name: "OpenAI Server",
+                    provider: "openai",
+                    models: ["gpt-4o", "gpt-4o-mini"],
+                    default: true,
+                },
+            ],
+        }
+        process.env.AI_MODELS_CONFIG = JSON.stringify(config)
+        process.env.AI_MODELS_CONFIG_PATH = "" // Clear file path
+
+        const models = await loadFlattenedServerModels()
+
+        expect(models.length).toBe(2)
+
+        const defaults = models.filter((m) => m.isDefault)
+        expect(defaults.length).toBe(1)
+
+        const defaultModel = defaults[0]
+        expect(defaultModel.provider).toBe("openai")
+        expect(defaultModel.modelId).toBe("gpt-4o") // First model of default provider
+    })
+})