feat: add local audio test mode

src/audio/guild-voice-session.ts

@@ -8,8 +8,10 @@ import {
   NoSubscriberBehavior,
   VoiceConnectionStatus,
   createAudioPlayer,
+  createAudioResource,
   entersState,
   joinVoiceChannel,
+  StreamType,
   type AudioPlayer,
   type AudioReceiveStream,
   type VoiceConnection,
@@ -21,7 +23,7 @@ import { Logger } from "../logger.js";
 import { float32ToPcm16Buffer, int16ArrayToFloat32, Stereo48kToMono16kDownsampler, takeFrame } from "./pcm.js";
 import { ConversationMemory, type UserUtterance } from "../services/conversation.js";
 import { ElevenLabsSttService } from "../services/elevenlabs-stt.js";
-import { ElevenLabsTtsService, type PreparedSpeechPlayback } from "../services/elevenlabs-tts.js";
+import { ElevenLabsTtsService, type PreparedSpeechAudio } from "../services/elevenlabs-tts.js";
 import { OpenAiLlmService } from "../services/openai-llm.js";
 
 interface GuildVoiceSessionOptions {
@@ -167,7 +169,7 @@ export class GuildVoiceSession extends EventEmitter {
 
   private draining = false;
   private currentAbortController: AbortController | null = null;
-  private currentPlayback: PreparedSpeechPlayback | null = null;
+  private currentPlayback: PreparedSpeechAudio | null = null;
   private textChannelId?: string;
 
   private constructor(private readonly options: GuildVoiceSessionOptions) {
@@ -415,7 +417,9 @@ export class GuildVoiceSession extends EventEmitter {
         }
 
         try {
-          const resource = this.currentPlayback.resource;
+          const resource = createAudioResource(this.currentPlayback.stream, {
+            inputType: StreamType.Raw,
+          });
           this.player.play(resource);
 
           await entersState(this.player, AudioPlayerStatus.Playing, 20_000).catch(() => null);