Feat: Frontend-Overhaul E — Lane-Editor, Routing-Policy (hot-reload) & Latenz-Karte

Teil 1: VoiceLatencyCard auf dem Dashboard (GET /api/voice/metrics, C2) —
zeigt STT/Vision/Chat-TTFB/TTS mit p50/p95/last + count.

Teil 2: UI-editierbare Routing-Policy. Neuer routing_policy.py (hot-reload JSON
unter MODELS_DIR/mc2-routing.json, Env=Defaults, atomarer Write, Validierung).
router_logic, gateway_proxy und gateway.routing_summary lesen jetzt live via
load_policy(); routing_summary ist lane-bewusst (chat/coding statt altem auto).
Neue Endpoints GET/PUT /api/routing/policy.

Teil 3: LaneEditor.tsx als ZONE im Cockpit (chat/coding-Aliase + Schwellen +
fast_no_think, Speichern/Default-je-Feld); Gateway-Node zeigt die Lanes.

Verifiziert: npm run build (tsc strict) clean, FastAPI TestClient (GET/PUT,
Validierung, Persistenz, Hot-reload durch die API), venv-Smoke (Routing).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

frontend/src/lib/api.ts

@@ -191,11 +191,50 @@ export interface RoutingResp {
   endpoint?: string
   heavy_threshold_chars?: number
   routes: { name: string; target: string }[]
+  lanes?: { name: string; target: string; threshold_chars?: number; escalate_chars?: number; aka?: string }[]
   fallbacks: Record<string, string[]>[]
   context_window_fallbacks: Record<string, string[]>[]
   gateway_reachable: boolean
 }
 
+// UI-editierbare Routing-Policy (GET/PUT /api/routing/policy). Aliase + Zeichen-Schwellen
+// hinter den Lanes; hot-reload im Backend (kein Restart).
+export interface RoutingPolicy {
+  fast: string
+  heavy: string
+  coder: string
+  coder_lite: string
+  heavy_chars: number
+  coding_escalate_chars: number
+  fast_no_think: boolean
+}
+export interface RoutingPolicyField {
+  key: keyof RoutingPolicy
+  label: string
+  type: "str" | "int" | "bool"
+  min?: number
+  max?: number
+}
+export interface RoutingPolicyMeta {
+  policy: RoutingPolicy
+  defaults: RoutingPolicy
+  fields: RoutingPolicyField[]
+}
+export const getRoutingPolicy = () => api<RoutingPolicyMeta>("/api/routing/policy")
+export const updateRoutingPolicy = (patch: Partial<RoutingPolicy>) =>
+  api<{ policy: RoutingPolicy }>("/api/routing/policy", { method: "PUT", body: JSON.stringify(patch) })
+
+// Per-Stage-Latenz der Voice/Lucy-Pipeline (GET /api/voice/metrics, C2).
+// Schlüssel = Stufe (stt|vision|chat_ttfb|tts), Wert = rollende Statistik.
+export interface VoiceStageStat {
+  count: number
+  avg_ms?: number
+  p50_ms?: number
+  p95_ms?: number
+  last_ms?: number
+}
+export type VoiceMetrics = Record<string, VoiceStageStat>
+
 export interface GitInfo {
   hash: string
   date: string

frontend/src/lib/queries.ts

@@ -18,10 +18,12 @@ import {
   type MemoryGraph,
   type ModelsResp,
   type RoutingResp,
+  type RoutingPolicyMeta,
   type ServicesResp,
   type SystemStatus,
   type TokenStats,
   type UpdatesResp,
+  type VoiceMetrics,
 } from "./api"
 
 // Zentrale Query-Keys (eine Quelle der Wahrheit für invalidate).
@@ -32,6 +34,8 @@ export const qk = {
   models: ["models"] as const,
   groups: ["groups"] as const,
   routing: ["routing"] as const,
+  routingPolicy: ["routing-policy"] as const,
+  voiceMetrics: ["voice-metrics"] as const,
   jobs: ["jobs"] as const,
   tokenStats: ["token-stats"] as const,
   agentStatus: ["agent-status"] as const,
@@ -70,6 +74,12 @@ export const useGroups = (refetchInterval = 8_000) =>
 export const useRouting = (refetchInterval = 4_000) =>
   useQuery({ queryKey: qk.routing, queryFn: () => api<RoutingResp>("/api/routing"), refetchInterval })
 
+export const useVoiceMetrics = (refetchInterval = 5_000) =>
+  useQuery({ queryKey: qk.voiceMetrics, queryFn: () => api<VoiceMetrics>("/api/voice/metrics"), refetchInterval })
+
+export const useRoutingPolicy = () =>
+  useQuery({ queryKey: qk.routingPolicy, queryFn: () => api<RoutingPolicyMeta>("/api/routing/policy") })
+
 export const useJobs = (refetchInterval = 2_000) =>
   useQuery({
     queryKey: qk.jobs,