feat(tts): add MiniMax TTS provider support
- Add DEFAULT_MINIMAX_MODEL constant - Add minimax to TELEGRAM_OUTPUT, DEFAULT_OUTPUT, TELEPHONY_OUTPUT - Add minimax config to ResolvedTtsConfig interface - Add minimax case in resolveTtsApiKey() and isTtsProviderConfigured() - Implement minimaxTTS() function with hex/URL audio handling - Add minimax branch in textToSpeech() and textToSpeechTelephony()
This commit is contained in:
leolele
parent
8beea7cd6e
commit
82f40f100c
154
src/tts/tts.ts
154
src/tts/tts.ts
@ -51,6 +51,7 @@ const DEFAULT_OPENAI_VOICE = "alloy";
|
||||
const DEFAULT_EDGE_VOICE = "en-US-MichelleNeural";
|
||||
const DEFAULT_EDGE_LANG = "en-US";
|
||||
const DEFAULT_EDGE_OUTPUT_FORMAT = "audio-24khz-48kbitrate-mono-mp3";
|
||||
const DEFAULT_MINIMAX_MODEL = "speech-2.8-turbo";
|
||||
|
||||
const DEFAULT_ELEVENLABS_VOICE_SETTINGS = {
|
||||
stability: 0.5,
|
||||
@ -65,13 +66,15 @@ const TELEGRAM_OUTPUT = {
|
||||
// ElevenLabs output formats use codec_sample_rate_bitrate naming.
|
||||
// Opus @ 48kHz/64kbps is a good voice-note tradeoff for Telegram.
|
||||
elevenlabs: "opus_48000_64",
|
||||
extension: ".opus",
|
||||
minimax: "mp3",
|
||||
extension: ".mp3",
|
||||
voiceCompatible: true,
|
||||
};
|
||||
|
||||
const DEFAULT_OUTPUT = {
|
||||
openai: "mp3" as const,
|
||||
elevenlabs: "mp3_44100_128",
|
||||
minimax: "mp3",
|
||||
extension: ".mp3",
|
||||
voiceCompatible: false,
|
||||
};
|
||||
@ -79,6 +82,7 @@ const DEFAULT_OUTPUT = {
|
||||
const TELEPHONY_OUTPUT = {
|
||||
openai: { format: "pcm" as const, sampleRate: 24000 },
|
||||
elevenlabs: { format: "pcm_22050", sampleRate: 22050 },
|
||||
minimax: { format: "mp3", sampleRate: 32000 },
|
||||
};
|
||||
|
||||
const TTS_AUTO_MODES = new Set<TtsAutoMode>(["off", "always", "inbound", "tagged"]);
|
||||
@ -124,6 +128,11 @@ export type ResolvedTtsConfig = {
|
||||
proxy?: string;
|
||||
timeoutMs?: number;
|
||||
};
|
||||
minimax: {
|
||||
apiKey?: string;
|
||||
model: string;
|
||||
baseUrl: string;
|
||||
};
|
||||
prefsPath?: string;
|
||||
maxTextLength: number;
|
||||
timeoutMs: number;
|
||||
@ -296,6 +305,11 @@ export function resolveTtsConfig(cfg: MoltbotConfig): ResolvedTtsConfig {
|
||||
proxy: raw.edge?.proxy?.trim() || undefined,
|
||||
timeoutMs: raw.edge?.timeoutMs,
|
||||
},
|
||||
minimax: {
|
||||
apiKey: raw.minimax?.apiKey,
|
||||
model: raw.minimax?.model ?? DEFAULT_MINIMAX_MODEL,
|
||||
baseUrl: raw.minimax?.baseUrl?.trim() || "https://api.minimaxi.com",
|
||||
},
|
||||
prefsPath: raw.prefsPath,
|
||||
maxTextLength: raw.maxTextLength ?? DEFAULT_MAX_TEXT_LENGTH,
|
||||
timeoutMs: raw.timeoutMs ?? DEFAULT_TIMEOUT_MS,
|
||||
@ -474,10 +488,13 @@ export function resolveTtsApiKey(
|
||||
if (provider === "openai") {
|
||||
return config.openai.apiKey || process.env.OPENAI_API_KEY;
|
||||
}
|
||||
if (provider === "minimax") {
|
||||
return config.minimax.apiKey || process.env.MINIMAX_API_KEY;
|
||||
}
|
||||
return undefined;
|
||||
}
|
||||
|
||||
export const TTS_PROVIDERS = ["openai", "elevenlabs", "edge"] as const;
|
||||
export const TTS_PROVIDERS = ["openai", "elevenlabs", "edge", "minimax"] as const;
|
||||
|
||||
export function resolveTtsProviderOrder(primary: TtsProvider): TtsProvider[] {
|
||||
return [primary, ...TTS_PROVIDERS.filter((provider) => provider !== primary)];
|
||||
@ -485,6 +502,9 @@ export function resolveTtsProviderOrder(primary: TtsProvider): TtsProvider[] {
|
||||
|
||||
export function isTtsProviderConfigured(config: ResolvedTtsConfig, provider: TtsProvider): boolean {
|
||||
if (provider === "edge") return config.edge.enabled;
|
||||
if (provider === "minimax") {
|
||||
return Boolean(resolveTtsApiKey(config, provider));
|
||||
}
|
||||
return Boolean(resolveTtsApiKey(config, provider));
|
||||
}
|
||||
|
||||
@ -1044,6 +1064,99 @@ async function openaiTTS(params: {
|
||||
}
|
||||
}
|
||||
|
||||
async function minimaxTTS(params: {
|
||||
text: string;
|
||||
apiKey: string;
|
||||
model: string;
|
||||
baseUrl: string;
|
||||
outputFormat: string;
|
||||
timeoutMs: number;
|
||||
}): Promise<Buffer> {
|
||||
const { text, apiKey, model, baseUrl, outputFormat, timeoutMs } = params;
|
||||
|
||||
const controller = new AbortController();
|
||||
const timeout = setTimeout(() => controller.abort(), timeoutMs);
|
||||
|
||||
try {
|
||||
const url = `${baseUrl.replace(/\/+$/, "")}/v1/t2a_v2`;
|
||||
|
||||
// Map output format to MiniMax audio format
|
||||
let audioFormat = "mp3";
|
||||
if (outputFormat.includes("opus") || outputFormat.includes("ogg")) {
|
||||
audioFormat = "mp3"; // MiniMax doesn't support opus directly
|
||||
} else if (outputFormat.includes("wav") || outputFormat.includes("pcm")) {
|
||||
audioFormat = "wav";
|
||||
} else if (outputFormat.includes("flac")) {
|
||||
audioFormat = "flac";
|
||||
}
|
||||
|
||||
const response = await fetch(url, {
|
||||
method: "POST",
|
||||
headers: {
|
||||
Authorization: `Bearer ${apiKey}`,
|
||||
"Content-Type": "application/json",
|
||||
},
|
||||
body: JSON.stringify({
|
||||
model,
|
||||
text,
|
||||
stream: false,
|
||||
voice_setting: {
|
||||
voice_id: "female-shaonv",
|
||||
speed: 1,
|
||||
vol: 1,
|
||||
pitch: 0,
|
||||
emotion: "neutral",
|
||||
},
|
||||
audio_setting: {
|
||||
sample_rate: 32000,
|
||||
bitrate: 128000,
|
||||
format: audioFormat,
|
||||
channel: 1,
|
||||
},
|
||||
output_format: outputFormat === "url" ? "url" : "hex",
|
||||
}),
|
||||
signal: controller.signal,
|
||||
});
|
||||
|
||||
if (!response.ok) {
|
||||
throw new Error(`MiniMax TTS API error (${response.status})`);
|
||||
}
|
||||
|
||||
const data = (await response.json()) as {
|
||||
data?: { audio_url?: string; audio?: string };
|
||||
base_resp?: { status_code?: number; status_msg?: string };
|
||||
};
|
||||
|
||||
if (data.base_resp?.status_code !== 0) {
|
||||
throw new Error(
|
||||
`MiniMax TTS API error: ${data.base_resp?.status_msg ?? "unknown error"}`,
|
||||
);
|
||||
}
|
||||
|
||||
// Handle hex format (default)
|
||||
if (outputFormat !== "url" && data.data?.audio) {
|
||||
return Buffer.from(data.data.audio, "hex");
|
||||
}
|
||||
|
||||
// Handle URL format
|
||||
if (data.data?.audio_url) {
|
||||
const audioResponse = await fetch(data.data.audio_url, {
|
||||
signal: controller.signal,
|
||||
});
|
||||
|
||||
if (!audioResponse.ok) {
|
||||
throw new Error(`Failed to fetch audio from URL (${audioResponse.status})`);
|
||||
}
|
||||
|
||||
return Buffer.from(await audioResponse.arrayBuffer());
|
||||
}
|
||||
|
||||
throw new Error("MiniMax TTS API did not return audio data");
|
||||
} finally {
|
||||
clearTimeout(timeout);
|
||||
}
|
||||
}
|
||||
|
||||
function inferEdgeExtension(outputFormat: string): string {
|
||||
const normalized = outputFormat.toLowerCase();
|
||||
if (normalized.includes("webm")) return ".webm";
|
||||
@ -1203,6 +1316,15 @@ export async function textToSpeech(params: {
|
||||
voiceSettings,
|
||||
timeoutMs: config.timeoutMs,
|
||||
});
|
||||
} else if (provider === "minimax") {
|
||||
audioBuffer = await minimaxTTS({
|
||||
text: params.text,
|
||||
apiKey,
|
||||
model: config.minimax.model,
|
||||
baseUrl: config.minimax.baseUrl,
|
||||
outputFormat: output.minimax,
|
||||
timeoutMs: config.timeoutMs,
|
||||
});
|
||||
} else {
|
||||
const openaiModelOverride = params.overrides?.openai?.model;
|
||||
const openaiVoiceOverride = params.overrides?.openai?.voice;
|
||||
@ -1228,7 +1350,12 @@ export async function textToSpeech(params: {
|
||||
audioPath,
|
||||
latencyMs,
|
||||
provider,
|
||||
outputFormat: provider === "openai" ? output.openai : output.elevenlabs,
|
||||
outputFormat:
|
||||
provider === "openai"
|
||||
? output.openai
|
||||
: provider === "minimax"
|
||||
? output.minimax
|
||||
: output.elevenlabs,
|
||||
voiceCompatible: output.voiceCompatible,
|
||||
};
|
||||
} catch (err) {
|
||||
@ -1307,6 +1434,27 @@ export async function textToSpeechTelephony(params: {
|
||||
};
|
||||
}
|
||||
|
||||
if (provider === "minimax") {
|
||||
const output = TELEPHONY_OUTPUT.minimax;
|
||||
const audioBuffer = await minimaxTTS({
|
||||
text: params.text,
|
||||
apiKey,
|
||||
model: config.minimax.model,
|
||||
baseUrl: config.minimax.baseUrl,
|
||||
outputFormat: output.format,
|
||||
timeoutMs: config.timeoutMs,
|
||||
});
|
||||
|
||||
return {
|
||||
success: true,
|
||||
audioBuffer,
|
||||
latencyMs: Date.now() - providerStart,
|
||||
provider,
|
||||
outputFormat: output.format,
|
||||
sampleRate: output.sampleRate,
|
||||
};
|
||||
}
|
||||
|
||||
const output = TELEPHONY_OUTPUT.openai;
|
||||
const audioBuffer = await openaiTTS({
|
||||
text: params.text,
|
||||
|
||||
Loading…
Reference in New Issue
Block a user