feat: add daily token limit with actual usage tracking (#171)

* feat: add daily token limit with actual usage tracking

- Add DAILY_TOKEN_LIMIT env var for configurable daily token limit
- Track actual tokens from Bedrock API response metadata (not estimates)
- Server sends inputTokens + cachedInputTokens + outputTokens via messageMetadata
- Client increments token count in onFinish callback with actual usage
- Add NaN guards to prevent corrupted localStorage values
- Add token limit toast notification with quota display
- Remove client-side token estimation (was blocking legitimate requests)
- Switch to js-tiktoken for client compatibility (pure JS, no WASM)

* feat: add TPM (tokens per minute) rate limiting

- Add 50k tokens/min client-side rate limit
- Track tokens per minute with automatic minute rollover
- Check TPM limit after daily limits pass
- Show toast when rate limit reached
- NaN guards for localStorage values

* feat: make TPM limit configurable via TPM_LIMIT env var

* chore: restore cache debug logs

* fix: prevent race condition in TPM tracking

checkTPMLimit was resetting TPM count to 0 when checking, which
overwrote the count saved by incrementTPMCount. Now checkTPMLimit
only reads and incrementTPMCount handles all writes.

* chore: improve TPM limit error message clarity

lib/token-counter.ts

@@ -1,21 +1,22 @@
 /**
- * Token counting utilities using Anthropic's tokenizer
+ * Token counting utilities using js-tiktoken
  *
- * This file is separate from system-prompts.ts because the @anthropic-ai/tokenizer
- * package uses WebAssembly which doesn't work well with Next.js server-side rendering.
- * Import this file only in scripts or client-side code, not in API routes.
+ * Uses cl100k_base encoding (GPT-4) which is close to Claude's tokenization.
+ * This is a pure JavaScript implementation, no WASM required.
  */
 
-import { countTokens } from "@anthropic-ai/tokenizer"
+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 using Anthropic's tokenizer
+ * 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 countTokens(text)
+    return encoder.encode(text).length
 }
 
 /**
@@ -28,8 +29,8 @@ export function getSystemPromptTokenCounts(): {
     extended: number
     additions: number
 } {
-    const defaultTokens = countTokens(DEFAULT_SYSTEM_PROMPT)
-    const extendedTokens = countTokens(EXTENDED_SYSTEM_PROMPT)
+    const defaultTokens = countTextTokens(DEFAULT_SYSTEM_PROMPT)
+    const extendedTokens = countTextTokens(EXTENDED_SYSTEM_PROMPT)
     return {
         default: defaultTokens,
         extended: extendedTokens,