Switch local TTS to Kokoro ONNX

src/config.ts

@@ -26,8 +26,10 @@ const envSchema = z.object({
   LOCAL_STT_DEVICE: z.string().min(1).default("auto"),
   LOCAL_STT_COMPUTE_TYPE: z.string().min(1).default("auto"),
   LOCAL_STT_BEAM_SIZE: z.coerce.number().int().min(1).max(8).default(1),
-  LOCAL_TTS_LANGUAGE: z.string().min(1).default("KR"),
-  LOCAL_TTS_SPEAKER: z.string().min(1).default("KR"),
+  LOCAL_TTS_MODEL_PATH: z.string().min(1).default(".local-ai/models/kokoro-v1.0.onnx"),
+  LOCAL_TTS_VOICES_PATH: z.string().min(1).default(".local-ai/models/voices-v1.0.bin"),
+  LOCAL_TTS_LANGUAGE: z.string().min(1).default("ko"),
+  LOCAL_TTS_SPEAKER: z.string().min(1).default("af_heart"),
   LOCAL_TTS_DEVICE: z.string().min(1).default("auto"),
   LOCAL_TTS_SPEED: z.coerce.number().min(0.8).max(1.6).default(1.12),
   BOT_DEFAULT_LANGUAGE: z.string().min(2).default("ko"),

src/discord-main.ts

@@ -16,7 +16,7 @@ import { GuildVoiceSession } from "./audio/guild-voice-session.js";
 import { type DiscordRuntimeConfig } from "./config.js";
 import { Logger } from "./logger.js";
 import { LocalFasterWhisperSttService } from "./services/local-stt.js";
-import { LocalMeloTtsService } from "./services/local-tts.js";
+import { LocalKokoroTtsService } from "./services/local-tts.js";
 import { OllamaLlmService } from "./services/ollama-llm.js";
 
 export async function runDiscordBot(config: DiscordRuntimeConfig, logger: Logger): Promise<void> {
@@ -38,7 +38,7 @@ export async function runDiscordBot(config: DiscordRuntimeConfig, logger: Logger
   });
 
   const stt = new LocalFasterWhisperSttService(config, logger);
-  const tts = new LocalMeloTtsService(config, logger);
+  const tts = new LocalKokoroTtsService(config, logger);
   const llm = new OllamaLlmService(config);
   const sessions = new Map<string, GuildVoiceSession>();
 

src/local-main.ts

@@ -6,7 +6,7 @@ import { Logger } from "./logger.js";
 import { LocalVoiceSession } from "./audio/local-voice-session.js";
 import { requireFfmpegPath } from "./audio/ffmpeg-path.js";
 import { LocalFasterWhisperSttService } from "./services/local-stt.js";
-import { LocalMeloTtsService } from "./services/local-tts.js";
+import { LocalKokoroTtsService } from "./services/local-tts.js";
 import { OllamaLlmService } from "./services/ollama-llm.js";
 
 export async function printLocalAudioDevices(): Promise<void> {
@@ -68,7 +68,7 @@ export async function printLocalAudioDevices(): Promise<void> {
 
 export async function runLocalAssistant(config: AssistantRuntimeConfig, logger: Logger): Promise<void> {
   const stt = new LocalFasterWhisperSttService(config, logger);
-  const tts = new LocalMeloTtsService(config, logger);
+  const tts = new LocalKokoroTtsService(config, logger);
   const llm = new OllamaLlmService(config);
 
   await stt.warmup();

src/python-runtime.ts

@@ -30,6 +30,14 @@ export function resolveLocalAiCachePath(config: AppConfig): string {
   return path.resolve(process.cwd(), config.LOCAL_AI_CACHE_DIR);
 }
 
+export function resolveLocalAiTtsModelPath(config: AppConfig): string {
+  return path.resolve(process.cwd(), config.LOCAL_TTS_MODEL_PATH);
+}
+
+export function resolveLocalAiTtsVoicesPath(config: AppConfig): string {
+  return path.resolve(process.cwd(), config.LOCAL_TTS_VOICES_PATH);
+}
+
 export function resolveVenvPythonPath(config: AppConfig): string {
   const venvPath = resolveLocalAiVenvPath(config);
   return process.platform === "win32"

src/services/local-tts.ts

@@ -7,12 +7,13 @@ import type { Logger } from "../logger.js";
 import { resolveFfmpegPath } from "../audio/ffmpeg-path.js";
 import { PythonJsonWorker } from "./python-json-worker.js";
 import type { PreparedSpeechAudio, TtsService } from "./tts.js";
+import { resolveLocalAiTtsModelPath, resolveLocalAiTtsVoicesPath } from "../python-runtime.js";
 
 interface SynthesizeResult {
   wav_base64?: string;
 }
 
-export class LocalMeloTtsService implements TtsService {
+export class LocalKokoroTtsService implements TtsService {
   private readonly worker: PythonJsonWorker;
 
   constructor(config: AssistantRuntimeConfig, logger: Logger) {
@@ -22,6 +23,8 @@ export class LocalMeloTtsService implements TtsService {
     }
 
     this.worker = new PythonJsonWorker(config, logger, "local_tts_worker.py", "local-tts", {
+      LOCAL_TTS_MODEL_PATH: resolveLocalAiTtsModelPath(config),
+      LOCAL_TTS_VOICES_PATH: resolveLocalAiTtsVoicesPath(config),
       LOCAL_TTS_LANGUAGE: config.LOCAL_TTS_LANGUAGE,
       LOCAL_TTS_SPEAKER: config.LOCAL_TTS_SPEAKER,
       LOCAL_TTS_DEVICE: config.LOCAL_TTS_DEVICE,

src/setup-local-ai.ts

@@ -1,10 +1,22 @@
 import { existsSync } from "node:fs";
-import { mkdir } from "node:fs/promises";
+import { mkdir, writeFile } from "node:fs/promises";
 import { spawn } from "node:child_process";
 import path from "node:path";
 
 import { loadConfig } from "./config.js";
-import { resolveLocalAiCachePath, resolveLocalAiVenvPath, resolvePythonLaunch, resolveVenvPythonPath } from "./python-runtime.js";
+import {
+  resolveLocalAiCachePath,
+  resolveLocalAiTtsModelPath,
+  resolveLocalAiTtsVoicesPath,
+  resolveLocalAiVenvPath,
+  resolvePythonLaunch,
+  resolveVenvPythonPath,
+} from "./python-runtime.js";
+
+const KOKORO_MODEL_URL =
+  "https://github.com/thewh1teagle/kokoro-onnx/releases/download/model-files-v1.0/kokoro-v1.0.onnx";
+const KOKORO_VOICES_URL =
+  "https://github.com/thewh1teagle/kokoro-onnx/releases/download/model-files-v1.0/voices-v1.0.bin";
 
 async function run(command: string, args: string[], extraEnv?: NodeJS.ProcessEnv): Promise<void> {
   await new Promise<void>((resolve, reject) => {
@@ -48,11 +60,28 @@ async function ensurePip(pythonBin: string, env: NodeJS.ProcessEnv): Promise<voi
   });
 }
 
+async function ensureDownload(url: string, filePath: string): Promise<void> {
+  if (existsSync(filePath)) {
+    return;
+  }
+
+  await mkdir(path.dirname(filePath), { recursive: true });
+  const response = await fetch(url);
+  if (!response.ok) {
+    throw new Error(`다운로드 실패: ${url} (${response.status})`);
+  }
+
+  const bytes = Buffer.from(await response.arrayBuffer());
+  await writeFile(filePath, bytes);
+}
+
 async function main(): Promise<void> {
   const config = loadConfig();
   const venvPath = resolveLocalAiVenvPath(config);
   const venvPython = resolveVenvPythonPath(config);
   const cachePath = resolveLocalAiCachePath(config);
+  const ttsModelPath = resolveLocalAiTtsModelPath(config);
+  const ttsVoicesPath = resolveLocalAiTtsVoicesPath(config);
   const requirementsPath = path.resolve(process.cwd(), "python", "requirements.txt");
   const baseEnv = {
     HF_HOME: cachePath,
@@ -77,6 +106,9 @@ async function main(): Promise<void> {
   console.log("로컬 AI 의존성 설치를 시작합니다.");
   await run(venvPython, ["-m", "pip", "install", "--upgrade", "pip", "setuptools", "wheel"], baseEnv);
   await run(venvPython, ["-m", "pip", "install", "-r", requirementsPath], baseEnv);
+  console.log("로컬 TTS 모델 파일을 확인합니다.");
+  await ensureDownload(KOKORO_MODEL_URL, ttsModelPath);
+  await ensureDownload(KOKORO_VOICES_URL, ttsVoicesPath);
 
   console.log("설치가 끝났습니다.");
   console.log("다음 순서:");