Merge remote-tracking branch 'origin/pr-484'

apps/aether-gateway/src/ai_serving/mod.rs

@@ -157,7 +157,9 @@ pub(crate) fn resolve_local_decision_execution_runtime_auth_context(
     decision: &GatewayControlDecision,
 ) -> Option<ExecutionRuntimeAuthContext> {
     resolve_decision_execution_runtime_auth_context(decision).filter(|auth_context| {
-        !auth_context.user_id.trim().is_empty() && !auth_context.api_key_id.trim().is_empty()
+        auth_context.access_allowed
+            && !auth_context.user_id.trim().is_empty()
+            && !auth_context.api_key_id.trim().is_empty()
     })
 }
 

apps/aether-gateway/src/control/auth/gate.rs

@@ -106,7 +106,7 @@ async fn balance_capacity_rejection(
     requested_model: Option<&str>,
     body: &Bytes,
 ) -> Result<Option<GatewayLocalAuthRejection>, GatewayError> {
-    if auth_context.api_key_is_standalone || auth_context.admin_bypass_limits {
+    if auth_context.api_key_is_standalone {
         return Ok(None);
     }
     if auth_context.local_rejection.is_some() {
@@ -816,6 +816,43 @@ mod tests {
         }
     }
 
+    #[tokio::test]
+    async fn admin_bypass_limits_does_not_skip_exhausted_daily_quota_capacity() {
+        let context = billing_context_with_pricing(
+            Some(json!({
+                "tiers": [{
+                    "up_to": null,
+                    "input_price_per_1m": 1.0,
+                    "output_price_per_1m": 2.0
+                }]
+            })),
+            None,
+            None,
+            None,
+        );
+        let state = state_with_quota_and_wallet(quota_availability(0.0, false), context);
+        let mut decision = decision_with_allowed_models(vec!["gpt-5".to_string()]);
+        if let Some(auth_context) = decision.auth_context.as_mut() {
+            auth_context.admin_bypass_limits = true;
+        }
+        let uri: Uri = "/v1/chat/completions".parse().expect("uri should parse");
+        let body = Bytes::from_static(
+            br#"{"model":"gpt-5","messages":[{"role":"user","content":"hi"}],"stream":true}"#,
+        );
+
+        let rejection =
+            request_model_local_rejection(&state, Some(&decision), &uri, &json_headers(), &body)
+                .await
+                .expect("quota rejection should resolve");
+
+        assert_eq!(
+            rejection,
+            Some(GatewayLocalAuthRejection::BalanceDenied {
+                remaining: Some(0.0),
+            })
+        );
+    }
+
     #[tokio::test]
     async fn positive_balance_still_denies_known_cost_above_available_capacity() {
         let context = billing_context_with_pricing(

apps/aether-gateway/src/control/auth/mod.rs

@@ -9,7 +9,8 @@ pub(crate) use gate::{
     request_model_local_rejection, should_buffer_request_for_local_auth,
     trusted_auth_local_rejection, GatewayLocalAuthRejection,
 };
-pub(super) use resolution::{resolve_control_decision_auth, ControlDecisionAuthResolution};
 pub(crate) use resolution::{
-    resolve_execution_runtime_auth_context, GatewayAdminPrincipalContext, GatewayControlAuthContext,
+    refresh_execution_runtime_auth_context, resolve_execution_runtime_auth_context,
+    GatewayAdminPrincipalContext, GatewayControlAuthContext,
 };
+pub(super) use resolution::{resolve_control_decision_auth, ControlDecisionAuthResolution};

apps/aether-gateway/src/control/auth/resolution.rs

@@ -433,7 +433,14 @@ pub(crate) async fn resolve_execution_runtime_auth_context(
     let _ = trace_id;
 
     if let Some(auth_context) = decision.auth_context.clone() {
-        return Ok(Some(auth_context));
+        return Ok(Some(
+            refresh_execution_runtime_auth_context(
+                state,
+                auth_context,
+                decision.auth_endpoint_signature.as_deref(),
+            )
+            .await?,
+        ));
     }
 
     let Some(auth_endpoint_signature) = decision.auth_endpoint_signature.as_deref() else {
@@ -445,7 +452,14 @@ pub(crate) async fn resolve_execution_runtime_auth_context(
     };
 
     if let Some(auth_context) = get_cached_auth_context(state, &cache_key) {
-        return Ok(Some(auth_context));
+        let refreshed = refresh_execution_runtime_auth_context(
+            state,
+            auth_context,
+            Some(auth_endpoint_signature),
+        )
+        .await?;
+        put_cached_auth_context(state, cache_key, refreshed.clone());
+        return Ok(Some(refreshed));
     }
 
     if let Some(auth_context) =
@@ -461,6 +475,56 @@ pub(crate) async fn resolve_execution_runtime_auth_context(
     Ok(None)
 }
 
+pub(crate) async fn refresh_execution_runtime_auth_context(
+    state: &AppState,
+    auth_context: GatewayControlAuthContext,
+    auth_endpoint_signature: Option<&str>,
+) -> Result<GatewayControlAuthContext, GatewayError> {
+    if auth_context.local_rejection.is_some() || !auth_context.access_allowed {
+        return Ok(auth_context);
+    }
+    let Some(auth_endpoint_signature) = auth_endpoint_signature
+        .map(str::trim)
+        .filter(|value| !value.is_empty())
+    else {
+        return Ok(auth_context);
+    };
+    if !state.has_auth_api_key_reader()
+        || auth_context.user_id.trim().is_empty()
+        || auth_context.api_key_id.trim().is_empty()
+    {
+        return Ok(auth_context);
+    }
+
+    let snapshot = state
+        .data
+        .read_auth_api_key_snapshot(
+            &auth_context.user_id,
+            &auth_context.api_key_id,
+            current_unix_secs(),
+        )
+        .await
+        .map_err(|err| GatewayError::Internal(err.to_string()))?;
+    let Some(snapshot) = snapshot else {
+        let mut denied = auth_context;
+        denied.access_allowed = false;
+        denied.local_rejection = Some(GatewayLocalAuthRejection::InvalidApiKey);
+        denied.balance_remaining = None;
+        return Ok(denied);
+    };
+
+    let wallet_access = resolve_wallet_auth_gate(state, &snapshot).await?;
+    Ok(build_data_backed_auth_context(
+        state,
+        snapshot,
+        auth_endpoint_signature,
+        Some(true),
+        auth_context.balance_remaining,
+        wallet_access,
+    )
+    .await)
+}
+
 fn put_cached_auth_context(
     state: &AppState,
     cache_key: String,
@@ -609,7 +673,7 @@ async fn build_data_backed_auth_context(
             .api_key_expires_at_unix_secs
             .is_some_and(|expires_at| expires_at < current_unix_secs());
     let locked_api_key = snapshot.api_key_is_locked && !snapshot.api_key_is_standalone;
-    let access_allowed = header_access_allowed
+    let key_access_allowed = header_access_allowed
         .map(|value| value && snapshot.currently_usable)
         .unwrap_or(snapshot.currently_usable);
     let wallet_remaining = wallet_access
@@ -656,7 +720,7 @@ async fn build_data_backed_auth_context(
         user_id: snapshot.user_id,
         api_key_id: snapshot.api_key_id,
         balance_remaining: wallet_remaining.or(balance_remaining),
-        access_allowed,
+        access_allowed: key_access_allowed && local_rejection.is_none(),
         user_rate_limit: snapshot.user_rate_limit,
         api_key_rate_limit: snapshot.api_key_rate_limit,
         api_key_is_standalone: snapshot.api_key_is_standalone,
@@ -835,13 +899,20 @@ mod tests {
         InMemoryAuthApiKeySnapshotRepository, StoredAuthApiKeySnapshot,
     };
     use aether_data::repository::provider_catalog::InMemoryProviderCatalogReadRepository;
+    use aether_data::repository::wallet::{
+        InMemoryWalletRepository, StoredWalletSnapshot, WalletReadRepository,
+    };
     use aether_data_contracts::repository::provider_catalog::{
         StoredProviderCatalogEndpoint, StoredProviderCatalogProvider,
     };
     use axum::http::{HeaderMap, Uri};
 
-    use super::{resolve_data_backed_auth_context, GatewayLocalAuthRejection};
+    use super::{
+        resolve_data_backed_auth_context, resolve_execution_runtime_auth_context,
+        GatewayLocalAuthRejection,
+    };
     use crate::control::auth::credentials::hash_api_key;
+    use crate::control::GatewayControlDecision;
     use crate::data::GatewayDataState;
     use crate::AppState;
 
@@ -946,6 +1017,154 @@ mod tests {
         assert_eq!(repository.touch_count("key-1"), 1);
     }
 
+    #[tokio::test]
+    async fn data_backed_auth_context_marks_wallet_denial_as_not_allowed() {
+        let api_key = "sk-test-empty-wallet";
+        let auth_repository = Arc::new(InMemoryAuthApiKeySnapshotRepository::seed(vec![(
+            Some(hash_api_key(api_key)),
+            sample_snapshot("key-empty-wallet", "user-empty-wallet"),
+        )]));
+        let wallet_repository = Arc::new(InMemoryWalletRepository::seed(vec![
+            StoredWalletSnapshot::new(
+                "wallet-empty".to_string(),
+                Some("user-empty-wallet".to_string()),
+                None,
+                0.0,
+                0.0,
+                "finite".to_string(),
+                "USD".to_string(),
+                "active".to_string(),
+                0.0,
+                0.0,
+                0.0,
+                0.0,
+                100,
+            )
+            .expect("wallet should build"),
+        ]));
+        let data =
+            GatewayDataState::with_auth_and_wallet_for_tests(auth_repository, wallet_repository);
+        let state = AppState::new()
+            .expect("state should build")
+            .with_data_state_for_tests(data);
+
+        let mut headers = HeaderMap::new();
+        headers.insert(
+            http::header::AUTHORIZATION,
+            format!("Bearer {api_key}").parse().unwrap(),
+        );
+
+        let auth_context = resolve_data_backed_auth_context(
+            &state,
+            &headers,
+            &uri("/v1/chat/completions"),
+            Some("openai:chat"),
+        )
+        .await
+        .expect("resolution should succeed")
+        .expect("auth context should exist");
+
+        assert_eq!(
+            auth_context.local_rejection,
+            Some(GatewayLocalAuthRejection::BalanceDenied {
+                remaining: Some(0.0),
+            })
+        );
+        assert!(!auth_context.access_allowed);
+    }
+
+    #[tokio::test]
+    async fn execution_runtime_auth_context_revalidates_cached_wallet_state() {
+        let api_key = "sk-test-runtime-wallet-cache";
+        let auth_repository = Arc::new(InMemoryAuthApiKeySnapshotRepository::seed(vec![(
+            Some(hash_api_key(api_key)),
+            sample_snapshot("key-runtime-wallet-cache", "user-runtime-wallet-cache"),
+        )]));
+        let wallet_repository = Arc::new(InMemoryWalletRepository::seed(vec![
+            StoredWalletSnapshot::new(
+                "wallet-runtime-cache".to_string(),
+                Some("user-runtime-wallet-cache".to_string()),
+                None,
+                10.0,
+                0.0,
+                "finite".to_string(),
+                "USD".to_string(),
+                "active".to_string(),
+                10.0,
+                0.0,
+                0.0,
+                0.0,
+                100,
+            )
+            .expect("wallet should build"),
+        ]));
+        let data = GatewayDataState::with_auth_and_wallet_for_tests(
+            auth_repository,
+            Arc::clone(&wallet_repository),
+        );
+        let state = AppState::new()
+            .expect("state should build")
+            .with_data_state_for_tests(data);
+        let decision = GatewayControlDecision::synthetic(
+            "/v1/chat/completions",
+            Some("ai_public".to_string()),
+            Some("openai".to_string()),
+            Some("chat".to_string()),
+            Some("openai:chat".to_string()),
+        );
+        let mut headers = HeaderMap::new();
+        headers.insert("x-api-key", api_key.parse().unwrap());
+
+        let first = resolve_execution_runtime_auth_context(
+            &state,
+            &decision,
+            &headers,
+            &uri("/v1/chat/completions"),
+            "trace-runtime-wallet-cache",
+        )
+        .await
+        .expect("resolution should succeed")
+        .expect("auth context should exist");
+        assert!(first.access_allowed);
+
+        wallet_repository
+            .update_auth_user_wallet_snapshot(
+                "user-runtime-wallet-cache",
+                0.0,
+                0.0,
+                "finite",
+                "USD",
+                "active",
+                10.0,
+                10.0,
+                0.0,
+                0.0,
+                Some(101),
+            )
+            .await
+            .expect("wallet update should succeed")
+            .expect("wallet should exist");
+
+        let second = resolve_execution_runtime_auth_context(
+            &state,
+            &decision,
+            &headers,
+            &uri("/v1/chat/completions"),
+            "trace-runtime-wallet-cache",
+        )
+        .await
+        .expect("resolution should succeed")
+        .expect("auth context should exist");
+
+        assert_eq!(
+            second.local_rejection,
+            Some(GatewayLocalAuthRejection::BalanceDenied {
+                remaining: Some(0.0),
+            })
+        );
+        assert!(!second.access_allowed);
+    }
+
     #[tokio::test]
     async fn data_backed_auth_context_allows_provider_id_for_matching_provider_type() {
         let api_key = "sk-test-provider-id";

apps/aether-gateway/src/control/mod.rs

@@ -8,9 +8,10 @@ mod public;
 mod route;
 
 pub(crate) use auth::{
-    extract_requested_model, request_model_local_rejection, resolve_execution_runtime_auth_context,
+    extract_requested_model, refresh_execution_runtime_auth_context, request_model_local_rejection,
-    should_buffer_request_for_local_auth, trusted_auth_local_rejection,
+    resolve_execution_runtime_auth_context, should_buffer_request_for_local_auth,
-    GatewayAdminPrincipalContext, GatewayControlAuthContext, GatewayLocalAuthRejection,
+    trusted_auth_local_rejection, GatewayAdminPrincipalContext, GatewayControlAuthContext,
+    GatewayLocalAuthRejection,
 };
 pub(crate) use execute::{allows_control_execute_emergency, maybe_execute_via_control};
 pub(crate) use management_token_permissions::{

apps/aether-gateway/src/handlers/internal/gateway.rs

@@ -30,6 +30,25 @@ use axum::response::IntoResponse;
 use axum::Json;
 use serde_json::json;
 
+async fn apply_supplied_auth_context(
+    state: &AppState,
+    decision: &mut GatewayControlDecision,
+    auth_context: Option<crate::control::GatewayControlAuthContext>,
+) -> Result<bool, GatewayError> {
+    let Some(auth_context) = auth_context else {
+        return Ok(false);
+    };
+    let refreshed = crate::control::refresh_execution_runtime_auth_context(
+        state,
+        auth_context,
+        decision.auth_endpoint_signature.as_deref(),
+    )
+    .await?;
+    decision.local_auth_rejection = refreshed.local_rejection.clone();
+    decision.auth_context = Some(refreshed);
+    Ok(true)
+}
+
 pub(crate) async fn maybe_build_local_internal_proxy_response_impl(
     state: &AppState,
     request_context: &GatewayPublicRequestContext,
@@ -206,11 +225,8 @@ pub(crate) async fn maybe_build_local_internal_proxy_response_impl(
                         Json(build_internal_gateway_fallback_plan_payload(None)).into_response(),
                     ));
                 };
-                let provided_auth_context = payload.auth_context.is_some();
+                let provided_auth_context =
-                if let Some(auth_context) = payload.auth_context {
+                    apply_supplied_auth_context(state, &mut resolved, payload.auth_context).await?;
-                    resolved.auth_context = Some(auth_context);
-                    resolved.local_auth_rejection = None;
-                }
                 let auth_context = resolved.auth_context.as_ref();
                 if auth_context
                     .map(|value| !value.access_allowed)
@@ -308,11 +324,8 @@ pub(crate) async fn maybe_build_local_internal_proxy_response_impl(
                         Json(build_internal_gateway_fallback_plan_payload(None)).into_response(),
                     ));
                 };
-                let provided_auth_context = payload.auth_context.is_some();
+                let provided_auth_context =
-                if let Some(auth_context) = payload.auth_context {
+                    apply_supplied_auth_context(state, &mut resolved, payload.auth_context).await?;
-                    resolved.auth_context = Some(auth_context);
-                    resolved.local_auth_rejection = None;
-                }
                 let auth_context = resolved.auth_context.as_ref();
                 if auth_context
                     .map(|value| !value.access_allowed)
@@ -407,11 +420,8 @@ pub(crate) async fn maybe_build_local_internal_proxy_response_impl(
                 else {
                     return Ok(Some(build_internal_gateway_proxy_public_response()));
                 };
-                let provided_auth_context = payload.auth_context.is_some();
+                let provided_auth_context =
-                if let Some(auth_context) = payload.auth_context {
+                    apply_supplied_auth_context(state, &mut resolved, payload.auth_context).await?;
-                    resolved.auth_context = Some(auth_context);
-                    resolved.local_auth_rejection = None;
-                }
                 if let Some(mut planned) = api::maybe_build_sync_plan_payload(
                     state,
                     &parts,
@@ -474,11 +484,8 @@ pub(crate) async fn maybe_build_local_internal_proxy_response_impl(
                 else {
                     return Ok(Some(build_internal_gateway_proxy_public_response()));
                 };
-                let provided_auth_context = payload.auth_context.is_some();
+                let provided_auth_context =
-                if let Some(auth_context) = payload.auth_context {
+                    apply_supplied_auth_context(state, &mut resolved, payload.auth_context).await?;
-                    resolved.auth_context = Some(auth_context);
-                    resolved.local_auth_rejection = None;
-                }
                 if let Some(mut planned) = api::maybe_build_stream_plan_payload(
                     state,
                     &parts,
@@ -546,10 +553,7 @@ pub(crate) async fn maybe_build_local_internal_proxy_response_impl(
                 else {
                     return Ok(None);
                 };
-                if let Some(auth_context) = payload.auth_context {
+                apply_supplied_auth_context(state, &mut resolved, payload.auth_context).await?;
-                    resolved.auth_context = Some(auth_context);
-                    resolved.local_auth_rejection = None;
-                }
                 if let Some(plan_payload) = api::maybe_build_sync_plan_payload(
                     state,
                     &parts,
@@ -630,10 +634,7 @@ pub(crate) async fn maybe_build_local_internal_proxy_response_impl(
                 else {
                     return Ok(None);
                 };
-                if let Some(auth_context) = payload.auth_context {
+                apply_supplied_auth_context(state, &mut resolved, payload.auth_context).await?;
-                    resolved.auth_context = Some(auth_context);
-                    resolved.local_auth_rejection = None;
-                }
                 if let Some(plan_payload) = api::maybe_build_stream_plan_payload(
                     state,
                     &parts,

apps/aether-gateway/src/handlers/public/support/wallet.rs

@@ -30,10 +30,10 @@ mod refunds;
 use self::flow::handle_wallet_flow;
 pub(in crate::handlers::public::support) use self::reads::build_wallet_balance_payload_for_user;
 use self::reads::{
-    build_wallet_daily_usage_payload, build_wallet_payload, build_wallet_zero_today_entry,
+    build_wallet_daily_usage_payload, build_wallet_live_today_usage_payload_for_user,
-    handle_wallet_balance, handle_wallet_today_cost, handle_wallet_transactions,
+    build_wallet_payload, build_wallet_zero_today_entry, handle_wallet_balance,
-    parse_wallet_limit, parse_wallet_offset, wallet_fixed_offset, wallet_today_billing_date_string,
+    handle_wallet_today_cost, handle_wallet_transactions, parse_wallet_limit, parse_wallet_offset,
-    wallet_transaction_payload_from_record,
+    wallet_fixed_offset, wallet_transaction_payload_from_record,
 };
 pub(crate) use self::recharge::sanitize_wallet_gateway_response;
 use self::recharge::{

apps/aether-gateway/src/handlers/public/support/wallet/flow.rs

@@ -1,9 +1,10 @@
 use super::{
     build_auth_error_response, build_auth_json_response, build_wallet_daily_usage_payload,
-    build_wallet_payload, build_wallet_zero_today_entry, http, parse_wallet_limit,
+    build_wallet_live_today_usage_payload_for_user, build_wallet_payload,
-    parse_wallet_offset, resolve_authenticated_local_user, unix_secs_to_rfc3339,
+    build_wallet_zero_today_entry, http, parse_wallet_limit, parse_wallet_offset,
-    wallet_fixed_offset, wallet_today_billing_date_string, wallet_transaction_payload_from_record,
+    resolve_authenticated_local_user, unix_secs_to_rfc3339, wallet_fixed_offset,
-    AppState, Body, GatewayPublicRequestContext, Response, WALLET_LEGACY_TIMEZONE,
+    wallet_transaction_payload_from_record, AppState, Body, GatewayPublicRequestContext, Response,
+    WALLET_LEGACY_TIMEZONE,
 };
 use serde_json::json;
 
@@ -98,7 +99,17 @@ pub(super) async fn handle_wallet_flow(
         return build_auth_json_response(http::StatusCode::OK, payload, None);
     };
 
-    let mut today_entry = build_wallet_zero_today_entry();
+    let mut today_entry =
+        match build_wallet_live_today_usage_payload_for_user(state, &auth.user.id).await {
+            Ok(Some(today_usage)) => today_usage,
+            _ => build_wallet_zero_today_entry(),
+        };
+    if today_entry
+        .get("total_requests")
+        .and_then(serde_json::Value::as_u64)
+        .unwrap_or_default()
+        == 0
+    {
         if let Ok(Some(today_usage)) = state
             .find_wallet_today_usage(&wallet.id, WALLET_LEGACY_TIMEZONE)
             .await
@@ -125,6 +136,7 @@ pub(super) async fn handle_wallet_flow(
                 true,
             );
         }
+    }
 
     let fetch_size = offset.saturating_add(limit).min(5200);
     let (transactions, tx_total) = state

apps/aether-gateway/src/handlers/public/support/wallet/reads.rs

@@ -1,11 +1,11 @@
 use super::{
     build_auth_error_response, build_auth_json_response, build_auth_wallet_summary_payload, http,
     query_param_value, resolve_authenticated_local_user, unix_secs_to_rfc3339, AppState, Body,
-    GatewayError, GatewayPublicRequestContext, Response, WALLET_LEGACY_TIMEZONE,
+    GatewayPublicRequestContext, Response, WALLET_LEGACY_TIMEZONE,
 };
 use crate::handlers::shared::round_to;
 use aether_data_contracts::repository::usage::UsageSettledCostSummaryQuery;
-use chrono::Utc;
+use chrono::{TimeZone, Utc};
 use serde_json::json;
 
 const WALLET_TODAY_COST_UNAVAILABLE_DETAIL: &str = "钱包今日费用数据暂不可用";
@@ -143,6 +143,26 @@ pub(super) fn wallet_today_billing_date_string() -> String {
         .to_string()
 }
 
+fn wallet_today_usage_window() -> Result<(String, String, u64, u64), String> {
+    let offset = wallet_fixed_offset();
+    let today = Utc::now().with_timezone(&offset).date_naive();
+    let Some(local_start_naive) = today.and_hms_opt(0, 0, 0) else {
+        return Err("wallet today start is invalid".to_string());
+    };
+    let Some(local_start) = offset.from_local_datetime(&local_start_naive).single() else {
+        return Err("wallet today local start is ambiguous".to_string());
+    };
+    let local_end = local_start + chrono::Duration::days(1);
+    let start_unix_secs = local_start.timestamp().max(0) as u64;
+    let end_unix_secs = local_end.timestamp().max(0) as u64;
+    Ok((
+        today.to_string(),
+        WALLET_LEGACY_TIMEZONE.to_string(),
+        start_unix_secs,
+        end_unix_secs,
+    ))
+}
+
 pub(super) fn build_wallet_daily_usage_payload(
     id: Option<String>,
     date: String,
@@ -193,6 +213,43 @@ pub(super) fn build_wallet_zero_today_entry() -> serde_json::Value {
     )
 }
 
+pub(super) async fn build_wallet_live_today_usage_payload_for_user(
+    state: &AppState,
+    user_id: &str,
+) -> Result<Option<serde_json::Value>, String> {
+    if !state.has_usage_data_reader() {
+        return Ok(None);
+    }
+    let (date, timezone, start_unix_secs, end_unix_secs) = wallet_today_usage_window()?;
+    let summary = state
+        .summarize_usage_settled_cost(&UsageSettledCostSummaryQuery {
+            created_from_unix_secs: start_unix_secs,
+            created_until_unix_secs: end_unix_secs,
+            user_id: Some(user_id.to_string()),
+        })
+        .await
+        .map_err(|err| format!("wallet today cost lookup failed: {err:?}"))?;
+    Ok(Some(build_wallet_daily_usage_payload(
+        None,
+        date,
+        timezone,
+        summary.total_cost_usd,
+        summary.total_requests,
+        summary.input_tokens,
+        summary.output_tokens,
+        summary.cache_creation_tokens,
+        summary.cache_read_tokens,
+        summary
+            .first_finalized_at_unix_secs
+            .and_then(unix_secs_to_rfc3339),
+        summary
+            .last_finalized_at_unix_secs
+            .and_then(unix_secs_to_rfc3339),
+        Some(Utc::now().to_rfc3339()),
+        true,
+    )))
+}
+
 pub(super) fn wallet_transaction_payload_from_record(
     record: &aether_data::repository::wallet::StoredAdminWalletTransaction,
 ) -> serde_json::Value {
@@ -253,65 +310,17 @@ pub(super) async fn handle_wallet_today_cost(
         Ok(value) => value,
         Err(response) => return response,
     };
-    let today = Utc::now().date_naive();
+    match build_wallet_live_today_usage_payload_for_user(state, &auth.user.id).await {
-    let Some(start_of_day) = today.and_hms_opt(0, 0, 0) else {
+        Ok(Some(payload)) => build_auth_json_response(http::StatusCode::OK, payload, None),
-        return build_auth_error_response(
+        Ok(None) => build_auth_error_response(
-            http::StatusCode::INTERNAL_SERVER_ERROR,
+            http::StatusCode::SERVICE_UNAVAILABLE,
-            "wallet today start is invalid",
+            WALLET_TODAY_COST_UNAVAILABLE_DETAIL,
             false,
-        );
+        ),
-    };
+        Err(detail) => {
-    let start_unix_secs = u64::try_from(
+            build_auth_error_response(http::StatusCode::INTERNAL_SERVER_ERROR, detail, false)
-        chrono::DateTime::<chrono::Utc>::from_naive_utc_and_offset(start_of_day, chrono::Utc)
+        }
-            .timestamp(),
-    )
-    .unwrap_or_default();
-    let end_unix_secs = start_unix_secs.saturating_add(24 * 3600);
-
-    let summary = match state
-        .summarize_usage_settled_cost(&UsageSettledCostSummaryQuery {
-            created_from_unix_secs: start_unix_secs,
-            created_until_unix_secs: end_unix_secs,
-            user_id: Some(auth.user.id.clone()),
-        })
-        .await
-    {
-        Ok(value) => value,
-        Err(err) => {
-            return build_auth_error_response(
-                http::StatusCode::INTERNAL_SERVER_ERROR,
-                format!("wallet today cost lookup failed: {err:?}"),
-                false,
-            )
     }
-    };
-
-    let first_finalized_at = summary
-        .first_finalized_at_unix_secs
-        .and_then(unix_secs_to_rfc3339);
-    let last_finalized_at = summary
-        .last_finalized_at_unix_secs
-        .and_then(unix_secs_to_rfc3339);
-
-    build_auth_json_response(
-        http::StatusCode::OK,
-        json!({
-            "id": serde_json::Value::Null,
-            "date": today.to_string(),
-            "timezone": "UTC",
-            "total_cost": round_to(summary.total_cost_usd, 6),
-            "total_requests": summary.total_requests,
-            "input_tokens": summary.input_tokens,
-            "output_tokens": summary.output_tokens,
-            "cache_creation_tokens": summary.cache_creation_tokens,
-            "cache_read_tokens": summary.cache_read_tokens,
-            "first_finalized_at": first_finalized_at,
-            "last_finalized_at": last_finalized_at,
-            "aggregated_at": Utc::now().to_rfc3339(),
-            "is_today": true,
-        }),
-        None,
-    )
 }
 
 pub(super) async fn handle_wallet_transactions(

apps/aether-gateway/src/tests/frontdoor/internal.rs

@@ -11,6 +11,9 @@ use crate::tests::{
     EXECUTION_PATH_EXECUTION_RUNTIME_STREAM, EXECUTION_PATH_EXECUTION_RUNTIME_SYNC,
     EXECUTION_PATH_HEADER,
 };
+use aether_data::repository::billing::InMemoryBillingReadRepository;
+use aether_data::repository::usage::InMemoryUsageReadRepository;
+use aether_data::repository::wallet::{InMemoryWalletRepository, StoredWalletSnapshot};
 use base64::Engine as _;
 
 #[tokio::test]
@@ -1154,6 +1157,118 @@ async fn gateway_handles_internal_gateway_decision_sync_locally_with_supplied_au
     upstream_handle.abort();
 }
 
+#[tokio::test]
+async fn gateway_internal_decision_sync_revalidates_supplied_auth_context_wallet() {
+    let upstream_hits = Arc::new(Mutex::new(0usize));
+    let upstream_hits_clone = Arc::clone(&upstream_hits);
+    let upstream = Router::new().route(
+        "/{*path}",
+        any(move |_request: Request| {
+            let upstream_hits_inner = Arc::clone(&upstream_hits_clone);
+            async move {
+                *upstream_hits_inner.lock().expect("mutex should lock") += 1;
+                (StatusCode::OK, Body::from("proxied"))
+            }
+        }),
+    );
+
+    let auth_repository = Arc::new(InMemoryAuthApiKeySnapshotRepository::seed(vec![(
+        None,
+        unrestricted_models_snapshot("api-key-empty-wallet", "user-empty-wallet"),
+    )]));
+    let candidate_repository =
+        Arc::new(InMemoryMinimalCandidateSelectionReadRepository::seed(vec![
+            sample_models_candidate_row("provider-1", "openai", "openai:chat", "gpt-5", 10),
+        ]));
+    let provider_catalog_repository = Arc::new(InMemoryProviderCatalogReadRepository::seed(
+        vec![sample_provider("provider-1", "openai", 10)],
+        vec![sample_endpoint(
+            "endpoint-provider-1",
+            "provider-1",
+            "openai:chat",
+            "https://api.openai.example",
+        )],
+        vec![sample_key(
+            "key-provider-1",
+            "provider-1",
+            "openai:chat",
+            "sk-upstream-openai",
+        )],
+    ));
+    let request_candidate_repository = Arc::new(InMemoryRequestCandidateRepository::default());
+    let usage_repository = Arc::new(InMemoryUsageReadRepository::default());
+    let billing_repository = Arc::new(InMemoryBillingReadRepository::seed(Vec::new()));
+    let wallet_repository = Arc::new(InMemoryWalletRepository::seed(vec![
+        StoredWalletSnapshot::new(
+            "wallet-empty".to_string(),
+            Some("user-empty-wallet".to_string()),
+            None,
+            0.0,
+            0.0,
+            "finite".to_string(),
+            "USD".to_string(),
+            "active".to_string(),
+            0.0,
+            0.0,
+            0.0,
+            0.0,
+            100,
+        )
+        .expect("wallet should build"),
+    ]));
+
+    let (upstream_url, upstream_handle) = start_server(upstream).await;
+    let gateway = build_router_with_state(
+        AppState::new()
+            .expect("gateway should build")
+            .with_data_state_for_tests(
+                crate::data::GatewayDataState::with_auth_candidate_selection_provider_catalog_request_candidates_usage_billing_and_wallet_for_tests(
+                    auth_repository,
+                    candidate_repository,
+                    provider_catalog_repository,
+                    request_candidate_repository,
+                    usage_repository,
+                    billing_repository,
+                    wallet_repository,
+                    DEVELOPMENT_ENCRYPTION_KEY,
+                ),
+            ),
+    );
+    let (gateway_url, gateway_handle) = start_server(gateway).await;
+
+    let response = reqwest::Client::new()
+        .post(format!("{gateway_url}/api/internal/gateway/decision-sync"))
+        .json(&json!({
+            "trace_id": "trace-internal-decision-sync-empty-wallet",
+            "method": "POST",
+            "path": "/v1/chat/completions",
+            "headers": {
+                "content-type": "application/json",
+            },
+            "body_json": {
+                "model": "gpt-5",
+                "messages": [],
+            },
+            "auth_context": {
+                "user_id": "user-empty-wallet",
+                "api_key_id": "api-key-empty-wallet",
+                "access_allowed": true,
+            }
+        }))
+        .send()
+        .await
+        .expect("request should succeed");
+
+    assert_eq!(response.status(), StatusCode::OK);
+    let payload: serde_json::Value = response.json().await.expect("json body should parse");
+    assert_eq!(payload["action"], "fallback_plan");
+    assert_eq!(payload["auth_context"], serde_json::Value::Null);
+    assert_eq!(*upstream_hits.lock().expect("mutex should lock"), 0);
+
+    gateway_handle.abort();
+    upstream_handle.abort();
+}
+
 #[tokio::test]
 async fn gateway_returns_internal_gateway_decision_sync_fallback_with_resolved_auth_context() {
     let upstream_hits = Arc::new(Mutex::new(0usize));

apps/aether-gateway/src/tests/frontdoor/public_support.rs

@@ -4133,13 +4133,7 @@ async fn gateway_handles_wallet_balance_locally_without_proxying_upstream() {
 #[tokio::test]
 async fn gateway_handles_wallet_today_cost_locally_without_proxying_upstream() {
     let auth_now = Utc::now();
-    let usage_now = chrono::DateTime::<chrono::Utc>::from_naive_utc_and_offset(
+    let usage_now = auth_now - chrono::Duration::minutes(30);
-        auth_now
-            .date_naive()
-            .and_hms_opt(12, 0, 0)
-            .expect("midday should be valid"),
-        chrono::Utc,
-    );
     let user = sample_auth_user(auth_now);
     let access_token = build_test_auth_token(
         "access",
@@ -4211,6 +4205,72 @@ async fn gateway_handles_wallet_today_cost_locally_without_proxying_upstream() {
     upstream_handle.abort();
 }
 
+#[tokio::test]
+async fn gateway_wallet_flow_today_entry_uses_live_settled_usage() {
+    let auth_now = Utc::now();
+    let user = sample_auth_user(auth_now);
+    let access_token = build_test_auth_token(
+        "access",
+        serde_json::Map::from_iter([
+            ("user_id".to_string(), json!(user.id)),
+            ("role".to_string(), json!(user.role)),
+            (
+                "created_at".to_string(),
+                json!(user.created_at.map(|value| value.to_rfc3339())),
+            ),
+            ("session_id".to_string(), json!("session-wallet-flow-today")),
+        ]),
+        auth_now + chrono::Duration::hours(1),
+    );
+    let usage_repository = Arc::new(InMemoryUsageReadRepository::seed(vec![
+        sample_user_usage_audit(
+            "usage-wallet-flow-today",
+            "req-wallet-flow-today",
+            "user-auth-1",
+            "gpt-4.1",
+            "OpenAI",
+            "completed",
+            auth_now - chrono::Duration::minutes(5),
+        ),
+    ]));
+    let (gateway_url, upstream_hits, gateway_handle, upstream_handle) =
+        start_auth_gateway_with_usage_state(
+            user,
+            sample_auth_wallet("user-auth-1", auth_now),
+            [sample_auth_session(
+                "user-auth-1",
+                "session-wallet-flow-today",
+                "device-wallet-flow-today",
+                "refresh-token-placeholder",
+                auth_now,
+            )],
+            usage_repository,
+        )
+        .await;
+
+    let response = reqwest::Client::new()
+        .get(format!("{gateway_url}/api/wallet/flow?limit=20&offset=0"))
+        .header("authorization", format!("Bearer {access_token}"))
+        .header("x-client-device-id", "device-wallet-flow-today")
+        .header("user-agent", "AetherTest/1.0")
+        .send()
+        .await
+        .expect("request should succeed");
+
+    assert_eq!(response.status(), StatusCode::OK);
+    let payload: serde_json::Value = response.json().await.expect("json body should parse");
+    assert_eq!(payload["today_entry"]["is_today"], true);
+    assert_eq!(payload["today_entry"]["timezone"], "Asia/Shanghai");
+    assert_eq!(payload["today_entry"]["total_requests"], 1);
+    assert_eq!(payload["today_entry"]["input_tokens"], 120);
+    assert_eq!(payload["today_entry"]["cache_read_tokens"], 15);
+    assert_eq!(payload["today_entry"]["total_cost"], 1.25);
+    assert_eq!(*upstream_hits.lock().expect("mutex should lock"), 0);
+
+    gateway_handle.abort();
+    upstream_handle.abort();
+}
+
 #[tokio::test]
 async fn gateway_returns_service_unavailable_for_wallet_today_cost_without_usage_reader() {
     let now = Utc::now();

apps/aether-gateway/src/wallet_runtime/access.rs

@@ -24,14 +24,9 @@ pub(crate) async fn resolve_wallet_auth_gate(
             auth_snapshot.api_key_is_standalone,
         )
         .await?;
-    let is_admin = wallet_auth_allows_admin_bypass(
-        &auth_snapshot.user_role,
-        auth_snapshot.api_key_is_standalone,
-    );
 
     let decision = match wallet.as_ref() {
-        Some(wallet) => map_wallet_snapshot(wallet).access_decision(is_admin),
+        Some(wallet) => map_wallet_snapshot(wallet).access_decision(false),
-        None if is_admin => WalletAccessDecision::allowed(None),
         None => WalletAccessDecision::wallet_unavailable(None),
     };
     if !auth_snapshot.api_key_is_standalone {
@@ -83,19 +78,50 @@ fn map_wallet_snapshot(snapshot: &StoredWalletSnapshot) -> WalletSnapshot {
     }
 }
 
-fn wallet_auth_allows_admin_bypass(user_role: &str, api_key_is_standalone: bool) -> bool {
-    user_role.eq_ignore_ascii_case("admin") && !api_key_is_standalone
-}
-
 #[cfg(test)]
 mod tests {
-    use aether_data::repository::wallet::StoredWalletSnapshot;
+    use std::sync::Arc;
+
+    use aether_data::repository::usage::InMemoryUsageReadRepository;
+    use aether_data::repository::wallet::{InMemoryWalletRepository, StoredWalletSnapshot};
+    use aether_data_contracts::repository::billing::{
+        BillingReadRepository, StoredBillingModelContext, UserDailyQuotaAvailabilityRecord,
+    };
+    use aether_data_contracts::DataLayerError;
     use aether_wallet::{WalletAccessFailure, WalletLimitMode, WalletSnapshot, WalletStatus};
+    use async_trait::async_trait;
 
     use super::{
-        local_rejection_from_wallet_access, map_wallet_snapshot, wallet_auth_allows_admin_bypass,
+        local_rejection_from_wallet_access, map_wallet_snapshot, resolve_wallet_auth_gate,
     };
     use crate::control::GatewayLocalAuthRejection;
+    use crate::data::auth::GatewayAuthApiKeySnapshot;
+    use crate::data::GatewayDataState;
+    use crate::AppState;
+
+    #[derive(Debug)]
+    struct FixedQuotaBillingReadRepository {
+        quota: Option<UserDailyQuotaAvailabilityRecord>,
+    }
+
+    #[async_trait]
+    impl BillingReadRepository for FixedQuotaBillingReadRepository {
+        async fn find_model_context(
+            &self,
+            _provider_id: &str,
+            _provider_api_key_id: Option<&str>,
+            _global_model_name: &str,
+        ) -> Result<Option<StoredBillingModelContext>, DataLayerError> {
+            Ok(None)
+        }
+
+        async fn find_user_daily_quota_availability(
+            &self,
+            _user_id: &str,
+        ) -> Result<Option<UserDailyQuotaAvailabilityRecord>, DataLayerError> {
+            Ok(self.quota.clone())
+        }
+    }
 
     #[test]
     fn maps_wallet_snapshot_and_derives_balance_denied() {
@@ -145,8 +171,143 @@ mod tests {
     }
 
     #[test]
-    fn standalone_key_never_uses_admin_wallet_bypass() {
+    fn admin_user_with_empty_finite_wallet_is_balance_denied() {
-        assert!(wallet_auth_allows_admin_bypass("admin", false));
+        let stored = StoredWalletSnapshot::new(
-        assert!(!wallet_auth_allows_admin_bypass("admin", true));
+            "wallet-1".to_string(),
+            Some("admin-1".to_string()),
+            None,
+            0.0,
+            0.0,
+            "finite".to_string(),
+            "USD".to_string(),
+            "active".to_string(),
+            0.0,
+            0.0,
+            0.0,
+            0.0,
+            100,
+        )
+        .expect("wallet should build");
+
+        let decision = map_wallet_snapshot(&stored).access_decision(true);
+
+        assert_eq!(decision.failure, Some(WalletAccessFailure::BalanceDenied));
+    }
+
+    #[tokio::test]
+    async fn ordinary_user_key_without_quota_denies_empty_wallet() {
+        let state = state_with_wallet_and_quota(empty_user_wallet(), None);
+        let auth_snapshot = ordinary_user_api_key_snapshot();
+
+        let decision = resolve_wallet_auth_gate(&state, &auth_snapshot)
+            .await
+            .expect("wallet gate should resolve")
+            .expect("wallet gate should return a decision");
+
+        assert!(!decision.allowed);
+        assert_eq!(decision.failure, Some(WalletAccessFailure::BalanceDenied));
+        assert_eq!(
+            local_rejection_from_wallet_access(&decision),
+            Some(GatewayLocalAuthRejection::BalanceDenied {
+                remaining: Some(0.0),
+            })
+        );
+    }
+
+    #[tokio::test]
+    async fn ordinary_user_key_with_remaining_quota_allows_empty_wallet() {
+        let state = state_with_wallet_and_quota(
+            empty_user_wallet(),
+            Some(quota_availability(10.0, 4.0, false)),
+        );
+        let auth_snapshot = ordinary_user_api_key_snapshot();
+
+        let decision = resolve_wallet_auth_gate(&state, &auth_snapshot)
+            .await
+            .expect("wallet gate should resolve")
+            .expect("wallet gate should return a decision");
+
+        assert!(decision.allowed);
+        assert_eq!(decision.failure, None);
+        assert_eq!(decision.remaining, Some(4.0));
+    }
+
+    fn state_with_wallet_and_quota(
+        wallet: StoredWalletSnapshot,
+        quota: Option<UserDailyQuotaAvailabilityRecord>,
+    ) -> AppState {
+        let usage_repository = Arc::new(InMemoryUsageReadRepository::default());
+        let billing_repository: Arc<dyn BillingReadRepository> =
+            Arc::new(FixedQuotaBillingReadRepository { quota });
+        let wallet_repository = Arc::new(InMemoryWalletRepository::seed(vec![wallet]));
+        let data = GatewayDataState::with_usage_billing_and_wallet_for_tests(
+            usage_repository,
+            billing_repository,
+            wallet_repository,
+        );
+        AppState::new()
+            .expect("state should build")
+            .with_data_state_for_tests(data)
+    }
+
+    fn empty_user_wallet() -> StoredWalletSnapshot {
+        StoredWalletSnapshot::new(
+            "wallet-user-1".to_string(),
+            Some("user-1".to_string()),
+            None,
+            0.0,
+            0.0,
+            "finite".to_string(),
+            "USD".to_string(),
+            "active".to_string(),
+            0.0,
+            0.0,
+            0.0,
+            0.0,
+            100,
+        )
+        .expect("wallet should build")
+    }
+
+    fn quota_availability(
+        total_quota_usd: f64,
+        remaining_usd: f64,
+        allow_wallet_overage: bool,
+    ) -> UserDailyQuotaAvailabilityRecord {
+        UserDailyQuotaAvailabilityRecord {
+            has_active_daily_quota: true,
+            total_quota_usd,
+            used_usd: total_quota_usd - remaining_usd,
+            remaining_usd,
+            allow_wallet_overage,
+        }
+    }
+
+    fn ordinary_user_api_key_snapshot() -> GatewayAuthApiKeySnapshot {
+        GatewayAuthApiKeySnapshot {
+            user_id: "user-1".to_string(),
+            username: "ordinary-user".to_string(),
+            email: Some("ordinary@example.com".to_string()),
+            user_role: "user".to_string(),
+            user_auth_source: "local".to_string(),
+            user_is_active: true,
+            user_is_deleted: false,
+            user_rate_limit: None,
+            user_allowed_providers: None,
+            user_allowed_api_formats: None,
+            user_allowed_models: None,
+            api_key_id: "api-key-1".to_string(),
+            api_key_name: Some("admin-created-key".to_string()),
+            api_key_is_active: true,
+            api_key_is_locked: false,
+            api_key_is_standalone: false,
+            api_key_rate_limit: None,
+            api_key_concurrent_limit: None,
+            api_key_expires_at_unix_secs: None,
+            api_key_allowed_providers: None,
+            api_key_allowed_api_formats: None,
+            api_key_allowed_models: None,
+            currently_usable: true,
+        }
     }
 }

crates/aether-data/src/repository/billing/sqlite.rs

@@ -961,7 +961,7 @@ ORDER BY expires_at ASC, created_at ASC, id ASC
             allow_wallet_overage &= grant.allow_wallet_overage;
             let used = sqlx::query_scalar::<_, f64>(
                 r#"
-SELECT COALESCE(SUM(amount_usd), 0)
+SELECT CAST(COALESCE(SUM(amount_usd), 0) AS REAL)
 FROM entitlement_usage_ledgers
 WHERE user_entitlement_id = ?
   AND usage_date = ?

crates/aether-data/src/repository/settlement/mysql.rs

@@ -464,6 +464,19 @@ FOR UPDATE
                 }
                 None => Some(0.0),
             };
+            if let Some(row) = wallet_row.as_ref() {
+                let wallet_id: String = row.try_get("id").map_sql_err()?;
+                let before_recharge: f64 = row.try_get("balance").map_sql_err()?;
+                let before_gift: f64 = row.try_get("gift_balance").map_sql_err()?;
+                let before_total = before_recharge + before_gift;
+                settlement.wallet_id = Some(wallet_id);
+                settlement.wallet_balance_before = Some(before_total);
+                settlement.wallet_balance_after = Some(before_total);
+                settlement.wallet_recharge_balance_before = Some(before_recharge);
+                settlement.wallet_recharge_balance_after = Some(before_recharge);
+                settlement.wallet_gift_balance_before = Some(before_gift);
+                settlement.wallet_gift_balance_after = Some(before_gift);
+            }
 
             let wallet_debit_cost_usd = if !api_key_is_standalone {
                 if let Some(user_id) = input.user_id.as_deref().filter(|value| !value.is_empty()) {

crates/aether-data/src/repository/settlement/postgres.rs

@@ -530,6 +530,20 @@ LIMIT 1
                             }
                             None => Some(0.0),
                         };
+                        if let Some(row) = wallet_row.as_ref() {
+                            let wallet_id: String = row.try_get("id").map_postgres_err()?;
+                            let before_recharge: f64 = row.try_get("balance").map_postgres_err()?;
+                            let before_gift: f64 =
+                                row.try_get("gift_balance").map_postgres_err()?;
+                            let before_total = before_recharge + before_gift;
+                            settlement.wallet_id = Some(wallet_id);
+                            settlement.wallet_balance_before = Some(before_total);
+                            settlement.wallet_balance_after = Some(before_total);
+                            settlement.wallet_recharge_balance_before = Some(before_recharge);
+                            settlement.wallet_recharge_balance_after = Some(before_recharge);
+                            settlement.wallet_gift_balance_before = Some(before_gift);
+                            settlement.wallet_gift_balance_after = Some(before_gift);
+                        }
 
                         let wallet_debit_cost_usd = if !api_key_is_standalone {
                             if let Some(user_id) =

crates/aether-data/src/repository/settlement/sqlite.rs

@@ -270,7 +270,7 @@ ORDER BY expires_at ASC, created_at ASC, id ASC
         allow_wallet_overage &= grant.allow_wallet_overage;
         let used = sqlx::query_scalar::<_, f64>(
             r#"
-SELECT COALESCE(SUM(amount_usd), 0)
+SELECT CAST(COALESCE(SUM(amount_usd), 0) AS REAL)
 FROM entitlement_usage_ledgers
 WHERE user_entitlement_id = ?
   AND usage_date = ?
@@ -475,6 +475,19 @@ LIMIT 1
                 }
                 None => Some(0.0),
             };
+            if let Some(row) = wallet_row.as_ref() {
+                let wallet_id: String = row.try_get("id").map_sql_err()?;
+                let before_recharge = sqlite_real(row, "balance")?;
+                let before_gift = sqlite_real(row, "gift_balance")?;
+                let before_total = before_recharge + before_gift;
+                settlement.wallet_id = Some(wallet_id);
+                settlement.wallet_balance_before = Some(before_total);
+                settlement.wallet_balance_after = Some(before_total);
+                settlement.wallet_recharge_balance_before = Some(before_recharge);
+                settlement.wallet_recharge_balance_after = Some(before_recharge);
+                settlement.wallet_gift_balance_before = Some(before_gift);
+                settlement.wallet_gift_balance_after = Some(before_gift);
+            }
 
             let wallet_debit_cost_usd = if !api_key_is_standalone {
                 if let Some(user_id) = input.user_id.as_deref().filter(|value| !value.is_empty()) {
@@ -802,6 +815,58 @@ mod tests {
         assert_eq!(wallet_total, 12.0);
     }
 
+    #[tokio::test]
+    async fn sqlite_repository_records_wallet_for_quota_covered_user_usage() {
+        let pool = sqlx::sqlite::SqlitePoolOptions::new()
+            .max_connections(1)
+            .connect("sqlite::memory:")
+            .await
+            .expect("sqlite pool should connect");
+        run_sqlite_migrations(&pool)
+            .await
+            .expect("sqlite migrations should run");
+        seed_quota_covered_settlement_rows(&pool).await;
+
+        let repository = SqliteSettlementRepository::new(pool.clone());
+        let settlement = repository
+            .settle_usage(UsageSettlementInput {
+                request_id: "request-quota-covered".to_string(),
+                user_id: Some("user-quota".to_string()),
+                api_key_id: Some("key-quota".to_string()),
+                api_key_is_standalone: false,
+                provider_id: None,
+                status: "completed".to_string(),
+                billing_status: "pending".to_string(),
+                total_cost_usd: 3.0,
+                actual_total_cost_usd: 2.0,
+                finalized_at_unix_secs: Some(1_260),
+            })
+            .await
+            .expect("settlement should run")
+            .expect("usage should exist");
+
+        assert_eq!(settlement.billing_status, "settled");
+        assert_eq!(settlement.wallet_id.as_deref(), Some("wallet-quota"));
+        assert_eq!(settlement.wallet_balance_before, Some(0.0));
+        assert_eq!(settlement.wallet_balance_after, Some(0.0));
+
+        let wallet_total: f64 = sqlx::query_scalar(
+            "SELECT balance + gift_balance FROM wallets WHERE id = 'wallet-quota'",
+        )
+        .fetch_one(&pool)
+        .await
+        .expect("wallet should load");
+        assert_eq!(wallet_total, 0.0);
+
+        let quota_used: f64 = sqlx::query_scalar(
+            "SELECT CAST(COALESCE(SUM(amount_usd), 0) AS REAL) FROM entitlement_usage_ledgers WHERE request_id = 'request-quota-covered'",
+        )
+        .fetch_one(&pool)
+        .await
+        .expect("quota ledger should load");
+        assert_eq!(quota_used, 3.0);
+    }
+
     async fn seed_settlement_rows(pool: &sqlx::SqlitePool) {
         sqlx::query(
             r#"
@@ -827,4 +892,62 @@ VALUES
         .await
         .expect("settlement rows should seed");
     }
+
+    async fn seed_quota_covered_settlement_rows(pool: &sqlx::SqlitePool) {
+        sqlx::query(
+            r#"
+INSERT INTO users (
+  id, username, email, role, auth_source, password_hash, is_active,
+  is_deleted, created_at, updated_at
+) VALUES (
+  'user-quota', 'quota-user', 'quota@example.com', 'user', 'local',
+  'hash', 1, 0, 1, 1
+);
+
+INSERT INTO wallets (
+  id, user_id, balance, gift_balance, limit_mode, created_at, updated_at
+) VALUES (
+  'wallet-quota', 'user-quota', 0.0, 0.0, 'finite', 1, 1
+);
+
+INSERT INTO "usage" (
+  request_id, user_id, api_key_id, status, billing_status,
+  total_cost_usd, actual_total_cost_usd
+) VALUES (
+  'request-quota-covered', 'user-quota', 'key-quota', 'completed',
+  'pending', 3.0, 2.0
+);
+
+INSERT INTO billing_plans (
+  id, title, price_amount, price_currency, duration_unit,
+  duration_value, entitlements_json, created_at, updated_at
+) VALUES (
+  'plan-quota', 'Quota Plan', 0.0, 'USD', 'month', 1,
+  '[{"type":"daily_quota","daily_quota_usd":10.0,"reset_timezone":"Asia/Shanghai","allow_wallet_overage":false}]',
+  1, 1
+);
+
+INSERT INTO payment_orders (
+  id, order_no, wallet_id, user_id, amount_usd, refunded_amount_usd,
+  refundable_amount_usd, payment_method, gateway_response, status, created_at
+) VALUES (
+  'order-quota', 'order-quota', 'wallet-quota', 'user-quota', 0.0, 0.0,
+  0.0, 'admin_manual', '{}', 'credited', 1
+);
+
+INSERT INTO user_plan_entitlements (
+  id, user_id, plan_id, payment_order_id, status, starts_at, expires_at,
+  entitlements_snapshot, created_at, updated_at
+) VALUES (
+  'entitlement-quota', 'user-quota', 'plan-quota', 'order-quota',
+  'active', 1, 9999999999,
+  '[{"type":"daily_quota","daily_quota_usd":10.0,"reset_timezone":"Asia/Shanghai","allow_wallet_overage":false}]',
+  1, 1
+);
+"#,
+        )
+        .execute(pool)
+        .await
+        .expect("quota settlement rows should seed");
+    }
 }

crates/aether-wallet/src/access.rs

@@ -63,10 +63,7 @@ impl WalletSnapshot {
         }
     }
 
-    pub fn access_decision(&self, is_admin: bool) -> WalletAccessDecision {
+    pub fn access_decision(&self, _is_admin: bool) -> WalletAccessDecision {
-        if is_admin {
-            return WalletAccessDecision::allowed(None);
-        }
         if self.status != WalletStatus::Active {
             return WalletAccessDecision::wallet_unavailable(self.balance_snapshot());
         }