refactor(server): centralize speech provider resolver typing

packages/server/src/server/agent/stt-manager.ts

@@ -1,5 +1,6 @@
 import type pino from "pino";
 import type { SpeechToTextProvider, TranscriptionResult } from "../speech/speech-provider.js";
+import { toResolver, type Resolvable } from "../speech/provider-resolver.js";
 import { maybePersistDebugAudio } from "./stt-debug.js";
 import { parsePcm16MonoWav, parsePcmRateFromFormat } from "../speech/audio.js";
 import { Pcm16MonoResampler } from "./pcm16-resampler.js";
@@ -44,11 +45,11 @@ export class STTManager {
   constructor(
     sessionId: string,
     logger: pino.Logger,
-    stt: SpeechToTextProvider | null | (() => SpeechToTextProvider | null)
+    stt: Resolvable<SpeechToTextProvider | null>
   ) {
     this.sessionId = sessionId;
     this.logger = logger.child({ module: "agent", component: "stt-manager", sessionId });
-    this.resolveStt = typeof stt === "function" ? stt : () => stt;
+    this.resolveStt = toResolver(stt);
   }
 
   /**

packages/server/src/server/agent/tts-manager.ts

@@ -1,6 +1,7 @@
 import type pino from "pino";
 import { v4 as uuidv4 } from "uuid";
 import type { TextToSpeechProvider } from "../speech/speech-provider.js";
+import { toResolver, type Resolvable } from "../speech/provider-resolver.js";
 import type { SessionOutboundMessage } from "../messages.js";
 
 interface PendingPlayback {
@@ -100,10 +101,10 @@ export class TTSManager {
   constructor(
     sessionId: string,
     logger: pino.Logger,
-    tts: TextToSpeechProvider | null | (() => TextToSpeechProvider | null)
+    tts: Resolvable<TextToSpeechProvider | null>
   ) {
     this.logger = logger.child({ module: "agent", component: "tts-manager", sessionId });
-    this.resolveTts = typeof tts === "function" ? tts : () => tts;
+    this.resolveTts = toResolver(tts);
   }
 
   /**

packages/server/src/server/dictation/dictation-stream-manager.ts

@@ -11,6 +11,7 @@ import type {
   SpeechToTextProvider,
   StreamingTranscriptionSession,
 } from "../speech/speech-provider.js";
+import { toResolver, type Resolvable } from "../speech/provider-resolver.js";
 import { parsePcmRateFromFormat, pcm16lePeakAbs } from "../speech/audio.js";
 
 const PCM_CHANNELS = 1;
@@ -126,19 +127,14 @@ export class DictationStreamManager {
     logger: pino.Logger;
     emit: (msg: DictationStreamOutboundMessage) => void;
     sessionId: string;
-    stt: SpeechToTextProvider | null | (() => SpeechToTextProvider | null);
+    stt: Resolvable<SpeechToTextProvider | null>;
     finalTimeoutMs?: number;
     autoCommitSeconds?: number;
   }) {
     this.logger = params.logger.child({ component: "dictation-stream-manager" });
     this.emit = params.emit;
     this.sessionId = params.sessionId;
-    if (typeof params.stt === "function") {
+    this.resolveStt = toResolver(params.stt);
-      this.resolveStt = params.stt;
-    } else {
-      const sttProvider = params.stt;
-      this.resolveStt = () => sttProvider;
-    }
     this.finalTimeoutMs = params.finalTimeoutMs ?? DEFAULT_DICTATION_FINAL_TIMEOUT_MS;
     this.autoCommitSeconds =
       params.autoCommitSeconds ??

packages/server/src/server/session.ts

@@ -130,6 +130,7 @@ import {
   listLocalSpeechModels,
   type LocalSpeechModelId,
 } from "./speech/providers/local/models.js";
+import type { Resolvable } from "./speech/provider-resolver.js";
 import type { SpeechReadinessSnapshot } from "./speech/speech-runtime.js";
 import type pino from "pino";
 
@@ -324,8 +325,8 @@ export type SessionOptions = {
   agentManager: AgentManager;
   agentStorage: AgentStorage;
   createAgentMcpTransport: AgentMcpTransportFactory;
-  stt: SpeechToTextProvider | null | (() => SpeechToTextProvider | null);
+  stt: Resolvable<SpeechToTextProvider | null>;
-  tts: TextToSpeechProvider | null | (() => TextToSpeechProvider | null);
+  tts: Resolvable<TextToSpeechProvider | null>;
   terminalManager: TerminalManager | null;
   voice?: {
     voiceAgentMcpStdio?: VoiceMcpStdioConfig | null;
@@ -340,7 +341,7 @@ export type SessionOptions = {
   };
   dictation?: {
     finalTimeoutMs?: number;
-    stt?: SpeechToTextProvider | null | (() => SpeechToTextProvider | null);
+    stt?: Resolvable<SpeechToTextProvider | null>;
     localModels?: {
       modelsDir: string;
       defaultModelIds: LocalSpeechModelId[];

packages/server/src/server/speech/provider-resolver.ts

@@ -0,0 +1,8 @@
+export type Resolvable<T> = T | (() => T);
+
+export function toResolver<T>(value: Resolvable<T>): () => T {
+  if (typeof value === "function") {
+    return value as () => T;
+  }
+  return () => value;
+}

packages/server/src/server/websocket-server.ts

@@ -26,6 +26,7 @@ import type { AgentProviderRuntimeSettingsMap } from "./agent/provider-launch-co
 import { PushTokenStore } from "./push/token-store.js";
 import { PushService } from "./push/push-service.js";
 import type { SpeechToTextProvider, TextToSpeechProvider } from "./speech/speech-provider.js";
+import type { Resolvable } from "./speech/provider-resolver.js";
 import type { SpeechReadinessSnapshot } from "./speech/speech-runtime.js";
 import type { LocalSpeechModelId } from "./speech/providers/local/models.js";
 import type {
@@ -94,18 +95,12 @@ export class VoiceAssistantWebSocketServer {
   private readonly pushTokenStore: PushTokenStore;
   private readonly pushService: PushService;
   private readonly createAgentMcpTransport: AgentMcpTransportFactory;
-  private readonly stt:
+  private readonly stt: Resolvable<SpeechToTextProvider | null>;
-    | SpeechToTextProvider
+  private readonly tts: Resolvable<TextToSpeechProvider | null>;
-    | null
-    | (() => SpeechToTextProvider | null);
-  private readonly tts:
-    | TextToSpeechProvider
-    | null
-    | (() => TextToSpeechProvider | null);
   private readonly terminalManager: TerminalManager | null;
   private readonly dictation: {
     finalTimeoutMs?: number;
-    stt?: SpeechToTextProvider | null | (() => SpeechToTextProvider | null);
+    stt?: Resolvable<SpeechToTextProvider | null>;
     localModels?: {
       modelsDir: string;
       defaultModelIds: LocalSpeechModelId[];
@@ -135,8 +130,8 @@ export class VoiceAssistantWebSocketServer {
     createAgentMcpTransport: AgentMcpTransportFactory,
     wsConfig: WebSocketServerConfig,
     speech?: {
-      stt: SpeechToTextProvider | null | (() => SpeechToTextProvider | null);
+      stt: Resolvable<SpeechToTextProvider | null>;
-      tts: TextToSpeechProvider | null | (() => TextToSpeechProvider | null);
+      tts: Resolvable<TextToSpeechProvider | null>;
     },
     terminalManager?: TerminalManager | null,
     voice?: {
@@ -146,7 +141,7 @@ export class VoiceAssistantWebSocketServer {
     },
     dictation?: {
       finalTimeoutMs?: number;
-      stt?: SpeechToTextProvider | null | (() => SpeechToTextProvider | null);
+      stt?: Resolvable<SpeechToTextProvider | null>;
       localModels?: {
         modelsDir: string;
         defaultModelIds: LocalSpeechModelId[];