mirror of
https://github.com/openclaw/openclaw.git
synced 2026-05-04 16:20:21 +00:00
refactor: plugin-own speech provider config
This commit is contained in:
Peter Steinberger
@@ -645,11 +645,10 @@ export class DiscordVoiceManager {
|
||||
cfg: this.params.cfg,
|
||||
override: this.params.discordConfig.voice?.tts,
|
||||
});
|
||||
const directive = parseTtsDirectives(
|
||||
replyText,
|
||||
ttsConfig.modelOverrides,
|
||||
ttsConfig.openai.baseUrl,
|
||||
);
|
||||
const directive = parseTtsDirectives(replyText, ttsConfig.modelOverrides, {
|
||||
cfg: ttsCfg,
|
||||
providerConfigs: ttsConfig.providerConfigs,
|
||||
});
|
||||
const speakText = directive.overrides.ttsText ?? directive.cleanedText.trim();
|
||||
if (!speakText) {
|
||||
logVoiceVerbose(
|
||||
|
||||
@@ -1,16 +1,309 @@
|
||||
import type { SpeechProviderPlugin } from "openclaw/plugin-sdk/core";
|
||||
import type { SpeechVoiceOption } from "openclaw/plugin-sdk/speech";
|
||||
import { normalizeResolvedSecretInputString } from "openclaw/plugin-sdk/secret-input";
|
||||
import type {
|
||||
SpeechDirectiveTokenParseContext,
|
||||
SpeechProviderConfig,
|
||||
SpeechProviderOverrides,
|
||||
SpeechProviderPlugin,
|
||||
SpeechVoiceOption,
|
||||
} from "openclaw/plugin-sdk/speech-core";
|
||||
import {
|
||||
normalizeApplyTextNormalization,
|
||||
normalizeLanguageCode,
|
||||
normalizeSeed,
|
||||
requireInRange,
|
||||
} from "openclaw/plugin-sdk/speech-core";
|
||||
import { elevenLabsTTS } from "./tts.js";
|
||||
|
||||
const DEFAULT_ELEVENLABS_BASE_URL = "https://api.elevenlabs.io";
|
||||
const DEFAULT_ELEVENLABS_VOICE_ID = "pMsXgVXv3BLzUgSXRplE";
|
||||
const DEFAULT_ELEVENLABS_MODEL_ID = "eleven_multilingual_v2";
|
||||
const DEFAULT_ELEVENLABS_VOICE_SETTINGS = {
|
||||
stability: 0.5,
|
||||
similarityBoost: 0.75,
|
||||
style: 0.0,
|
||||
useSpeakerBoost: true,
|
||||
speed: 1.0,
|
||||
};
|
||||
|
||||
const ELEVENLABS_TTS_MODELS = [
|
||||
"eleven_multilingual_v2",
|
||||
"eleven_turbo_v2_5",
|
||||
"eleven_monolingual_v1",
|
||||
] as const;
|
||||
|
||||
type ElevenLabsProviderConfig = {
|
||||
apiKey?: string;
|
||||
baseUrl: string;
|
||||
voiceId: string;
|
||||
modelId: string;
|
||||
seed?: number;
|
||||
applyTextNormalization?: "auto" | "on" | "off";
|
||||
languageCode?: string;
|
||||
voiceSettings: {
|
||||
stability: number;
|
||||
similarityBoost: number;
|
||||
style: number;
|
||||
useSpeakerBoost: boolean;
|
||||
speed: number;
|
||||
};
|
||||
};
|
||||
|
||||
function trimToUndefined(value: unknown): string | undefined {
|
||||
return typeof value === "string" && value.trim() ? value.trim() : undefined;
|
||||
}
|
||||
|
||||
function asNumber(value: unknown): number | undefined {
|
||||
return typeof value === "number" && Number.isFinite(value) ? value : undefined;
|
||||
}
|
||||
|
||||
function asBoolean(value: unknown): boolean | undefined {
|
||||
return typeof value === "boolean" ? value : undefined;
|
||||
}
|
||||
|
||||
function asObject(value: unknown): Record<string, unknown> | undefined {
|
||||
return typeof value === "object" && value !== null && !Array.isArray(value)
|
||||
? (value as Record<string, unknown>)
|
||||
: undefined;
|
||||
}
|
||||
|
||||
function parseBooleanValue(value: string): boolean | undefined {
|
||||
const normalized = value.trim().toLowerCase();
|
||||
if (["true", "1", "yes", "on"].includes(normalized)) {
|
||||
return true;
|
||||
}
|
||||
if (["false", "0", "no", "off"].includes(normalized)) {
|
||||
return false;
|
||||
}
|
||||
return undefined;
|
||||
}
|
||||
|
||||
function parseNumberValue(value: string): number | undefined {
|
||||
const parsed = Number.parseFloat(value);
|
||||
return Number.isFinite(parsed) ? parsed : undefined;
|
||||
}
|
||||
|
||||
export function isValidVoiceId(voiceId: string): boolean {
|
||||
return /^[a-zA-Z0-9]{10,40}$/.test(voiceId);
|
||||
}
|
||||
|
||||
function normalizeElevenLabsBaseUrl(baseUrl: string | undefined): string {
|
||||
const trimmed = baseUrl?.trim();
|
||||
return trimmed?.replace(/\/+$/, "") || "https://api.elevenlabs.io";
|
||||
return trimmed?.replace(/\/+$/, "") || DEFAULT_ELEVENLABS_BASE_URL;
|
||||
}
|
||||
|
||||
function normalizeElevenLabsProviderConfig(
|
||||
rawConfig: Record<string, unknown>,
|
||||
): ElevenLabsProviderConfig {
|
||||
const raw = asObject(rawConfig.elevenlabs);
|
||||
const rawVoiceSettings = asObject(raw?.voiceSettings);
|
||||
return {
|
||||
apiKey: normalizeResolvedSecretInputString({
|
||||
value: raw?.apiKey,
|
||||
path: "messages.tts.elevenlabs.apiKey",
|
||||
}),
|
||||
baseUrl: normalizeElevenLabsBaseUrl(trimToUndefined(raw?.baseUrl)),
|
||||
voiceId: trimToUndefined(raw?.voiceId) ?? DEFAULT_ELEVENLABS_VOICE_ID,
|
||||
modelId: trimToUndefined(raw?.modelId) ?? DEFAULT_ELEVENLABS_MODEL_ID,
|
||||
seed: asNumber(raw?.seed),
|
||||
applyTextNormalization: trimToUndefined(raw?.applyTextNormalization) as
|
||||
| "auto"
|
||||
| "on"
|
||||
| "off"
|
||||
| undefined,
|
||||
languageCode: trimToUndefined(raw?.languageCode),
|
||||
voiceSettings: {
|
||||
stability:
|
||||
asNumber(rawVoiceSettings?.stability) ?? DEFAULT_ELEVENLABS_VOICE_SETTINGS.stability,
|
||||
similarityBoost:
|
||||
asNumber(rawVoiceSettings?.similarityBoost) ??
|
||||
DEFAULT_ELEVENLABS_VOICE_SETTINGS.similarityBoost,
|
||||
style: asNumber(rawVoiceSettings?.style) ?? DEFAULT_ELEVENLABS_VOICE_SETTINGS.style,
|
||||
useSpeakerBoost:
|
||||
asBoolean(rawVoiceSettings?.useSpeakerBoost) ??
|
||||
DEFAULT_ELEVENLABS_VOICE_SETTINGS.useSpeakerBoost,
|
||||
speed: asNumber(rawVoiceSettings?.speed) ?? DEFAULT_ELEVENLABS_VOICE_SETTINGS.speed,
|
||||
},
|
||||
};
|
||||
}
|
||||
|
||||
function readElevenLabsProviderConfig(config: SpeechProviderConfig): ElevenLabsProviderConfig {
|
||||
const defaults = normalizeElevenLabsProviderConfig({});
|
||||
const voiceSettings = asObject(config.voiceSettings);
|
||||
return {
|
||||
apiKey: trimToUndefined(config.apiKey) ?? defaults.apiKey,
|
||||
baseUrl: normalizeElevenLabsBaseUrl(trimToUndefined(config.baseUrl) ?? defaults.baseUrl),
|
||||
voiceId: trimToUndefined(config.voiceId) ?? defaults.voiceId,
|
||||
modelId: trimToUndefined(config.modelId) ?? defaults.modelId,
|
||||
seed: asNumber(config.seed) ?? defaults.seed,
|
||||
applyTextNormalization:
|
||||
(trimToUndefined(config.applyTextNormalization) as "auto" | "on" | "off" | undefined) ??
|
||||
defaults.applyTextNormalization,
|
||||
languageCode: trimToUndefined(config.languageCode) ?? defaults.languageCode,
|
||||
voiceSettings: {
|
||||
stability: asNumber(voiceSettings?.stability) ?? defaults.voiceSettings.stability,
|
||||
similarityBoost:
|
||||
asNumber(voiceSettings?.similarityBoost) ?? defaults.voiceSettings.similarityBoost,
|
||||
style: asNumber(voiceSettings?.style) ?? defaults.voiceSettings.style,
|
||||
useSpeakerBoost:
|
||||
asBoolean(voiceSettings?.useSpeakerBoost) ?? defaults.voiceSettings.useSpeakerBoost,
|
||||
speed: asNumber(voiceSettings?.speed) ?? defaults.voiceSettings.speed,
|
||||
},
|
||||
};
|
||||
}
|
||||
|
||||
function mergeVoiceSettingsOverride(
|
||||
ctx: SpeechDirectiveTokenParseContext,
|
||||
next: Record<string, unknown>,
|
||||
): SpeechProviderOverrides {
|
||||
return {
|
||||
...(ctx.currentOverrides ?? {}),
|
||||
voiceSettings: {
|
||||
...(asObject(ctx.currentOverrides?.voiceSettings) ?? {}),
|
||||
...next,
|
||||
},
|
||||
};
|
||||
}
|
||||
|
||||
function parseDirectiveToken(ctx: SpeechDirectiveTokenParseContext) {
|
||||
try {
|
||||
switch (ctx.key) {
|
||||
case "voiceid":
|
||||
case "voice_id":
|
||||
case "elevenlabs_voice":
|
||||
case "elevenlabsvoice":
|
||||
if (!ctx.policy.allowVoice) {
|
||||
return { handled: true };
|
||||
}
|
||||
if (!isValidVoiceId(ctx.value)) {
|
||||
return { handled: true, warnings: [`invalid ElevenLabs voiceId "${ctx.value}"`] };
|
||||
}
|
||||
return {
|
||||
handled: true,
|
||||
overrides: { ...(ctx.currentOverrides ?? {}), voiceId: ctx.value },
|
||||
};
|
||||
case "model":
|
||||
case "modelid":
|
||||
case "model_id":
|
||||
case "elevenlabs_model":
|
||||
case "elevenlabsmodel":
|
||||
if (!ctx.policy.allowModelId) {
|
||||
return { handled: true };
|
||||
}
|
||||
return {
|
||||
handled: true,
|
||||
overrides: { ...(ctx.currentOverrides ?? {}), modelId: ctx.value },
|
||||
};
|
||||
case "stability": {
|
||||
if (!ctx.policy.allowVoiceSettings) {
|
||||
return { handled: true };
|
||||
}
|
||||
const value = parseNumberValue(ctx.value);
|
||||
if (value == null) {
|
||||
return { handled: true, warnings: ["invalid stability value"] };
|
||||
}
|
||||
requireInRange(value, 0, 1, "stability");
|
||||
return { handled: true, overrides: mergeVoiceSettingsOverride(ctx, { stability: value }) };
|
||||
}
|
||||
case "similarity":
|
||||
case "similarityboost":
|
||||
case "similarity_boost": {
|
||||
if (!ctx.policy.allowVoiceSettings) {
|
||||
return { handled: true };
|
||||
}
|
||||
const value = parseNumberValue(ctx.value);
|
||||
if (value == null) {
|
||||
return { handled: true, warnings: ["invalid similarityBoost value"] };
|
||||
}
|
||||
requireInRange(value, 0, 1, "similarityBoost");
|
||||
return {
|
||||
handled: true,
|
||||
overrides: mergeVoiceSettingsOverride(ctx, { similarityBoost: value }),
|
||||
};
|
||||
}
|
||||
case "style": {
|
||||
if (!ctx.policy.allowVoiceSettings) {
|
||||
return { handled: true };
|
||||
}
|
||||
const value = parseNumberValue(ctx.value);
|
||||
if (value == null) {
|
||||
return { handled: true, warnings: ["invalid style value"] };
|
||||
}
|
||||
requireInRange(value, 0, 1, "style");
|
||||
return { handled: true, overrides: mergeVoiceSettingsOverride(ctx, { style: value }) };
|
||||
}
|
||||
case "speed": {
|
||||
if (!ctx.policy.allowVoiceSettings) {
|
||||
return { handled: true };
|
||||
}
|
||||
const value = parseNumberValue(ctx.value);
|
||||
if (value == null) {
|
||||
return { handled: true, warnings: ["invalid speed value"] };
|
||||
}
|
||||
requireInRange(value, 0.5, 2, "speed");
|
||||
return { handled: true, overrides: mergeVoiceSettingsOverride(ctx, { speed: value }) };
|
||||
}
|
||||
case "speakerboost":
|
||||
case "speaker_boost":
|
||||
case "usespeakerboost":
|
||||
case "use_speaker_boost": {
|
||||
if (!ctx.policy.allowVoiceSettings) {
|
||||
return { handled: true };
|
||||
}
|
||||
const value = parseBooleanValue(ctx.value);
|
||||
if (value == null) {
|
||||
return { handled: true, warnings: ["invalid useSpeakerBoost value"] };
|
||||
}
|
||||
return {
|
||||
handled: true,
|
||||
overrides: mergeVoiceSettingsOverride(ctx, { useSpeakerBoost: value }),
|
||||
};
|
||||
}
|
||||
case "normalize":
|
||||
case "applytextnormalization":
|
||||
case "apply_text_normalization":
|
||||
if (!ctx.policy.allowNormalization) {
|
||||
return { handled: true };
|
||||
}
|
||||
return {
|
||||
handled: true,
|
||||
overrides: {
|
||||
...(ctx.currentOverrides ?? {}),
|
||||
applyTextNormalization: normalizeApplyTextNormalization(ctx.value),
|
||||
},
|
||||
};
|
||||
case "language":
|
||||
case "languagecode":
|
||||
case "language_code":
|
||||
if (!ctx.policy.allowNormalization) {
|
||||
return { handled: true };
|
||||
}
|
||||
return {
|
||||
handled: true,
|
||||
overrides: {
|
||||
...(ctx.currentOverrides ?? {}),
|
||||
languageCode: normalizeLanguageCode(ctx.value),
|
||||
},
|
||||
};
|
||||
case "seed":
|
||||
if (!ctx.policy.allowSeed) {
|
||||
return { handled: true };
|
||||
}
|
||||
return {
|
||||
handled: true,
|
||||
overrides: {
|
||||
...(ctx.currentOverrides ?? {}),
|
||||
seed: normalizeSeed(Number.parseInt(ctx.value, 10)),
|
||||
},
|
||||
};
|
||||
default:
|
||||
return { handled: false };
|
||||
}
|
||||
} catch (error) {
|
||||
return {
|
||||
handled: true,
|
||||
warnings: [error instanceof Error ? error.message : String(error)],
|
||||
};
|
||||
}
|
||||
}
|
||||
|
||||
export async function listElevenLabsVoices(params: {
|
||||
@@ -49,49 +342,164 @@ export function buildElevenLabsSpeechProvider(): SpeechProviderPlugin {
|
||||
return {
|
||||
id: "elevenlabs",
|
||||
label: "ElevenLabs",
|
||||
autoSelectOrder: 20,
|
||||
models: ELEVENLABS_TTS_MODELS,
|
||||
resolveConfig: ({ rawConfig }) => normalizeElevenLabsProviderConfig(rawConfig),
|
||||
parseDirectiveToken,
|
||||
resolveTalkConfig: ({ baseTtsConfig, talkProviderConfig }) => {
|
||||
const base = normalizeElevenLabsProviderConfig(baseTtsConfig);
|
||||
const talkVoiceSettings = asObject(talkProviderConfig.voiceSettings);
|
||||
return {
|
||||
...base,
|
||||
...(talkProviderConfig.apiKey === undefined
|
||||
? {}
|
||||
: {
|
||||
apiKey: normalizeResolvedSecretInputString({
|
||||
value: talkProviderConfig.apiKey,
|
||||
path: "talk.providers.elevenlabs.apiKey",
|
||||
}),
|
||||
}),
|
||||
...(trimToUndefined(talkProviderConfig.baseUrl) == null
|
||||
? {}
|
||||
: { baseUrl: normalizeElevenLabsBaseUrl(trimToUndefined(talkProviderConfig.baseUrl)) }),
|
||||
...(trimToUndefined(talkProviderConfig.voiceId) == null
|
||||
? {}
|
||||
: { voiceId: trimToUndefined(talkProviderConfig.voiceId) }),
|
||||
...(trimToUndefined(talkProviderConfig.modelId) == null
|
||||
? {}
|
||||
: { modelId: trimToUndefined(talkProviderConfig.modelId) }),
|
||||
...(asNumber(talkProviderConfig.seed) == null
|
||||
? {}
|
||||
: { seed: asNumber(talkProviderConfig.seed) }),
|
||||
...(trimToUndefined(talkProviderConfig.applyTextNormalization) == null
|
||||
? {}
|
||||
: {
|
||||
applyTextNormalization: normalizeApplyTextNormalization(
|
||||
trimToUndefined(talkProviderConfig.applyTextNormalization),
|
||||
),
|
||||
}),
|
||||
...(trimToUndefined(talkProviderConfig.languageCode) == null
|
||||
? {}
|
||||
: {
|
||||
languageCode: normalizeLanguageCode(trimToUndefined(talkProviderConfig.languageCode)),
|
||||
}),
|
||||
voiceSettings: {
|
||||
...base.voiceSettings,
|
||||
...(asNumber(talkVoiceSettings?.stability) == null
|
||||
? {}
|
||||
: { stability: asNumber(talkVoiceSettings?.stability) }),
|
||||
...(asNumber(talkVoiceSettings?.similarityBoost) == null
|
||||
? {}
|
||||
: { similarityBoost: asNumber(talkVoiceSettings?.similarityBoost) }),
|
||||
...(asNumber(talkVoiceSettings?.style) == null
|
||||
? {}
|
||||
: { style: asNumber(talkVoiceSettings?.style) }),
|
||||
...(asBoolean(talkVoiceSettings?.useSpeakerBoost) == null
|
||||
? {}
|
||||
: { useSpeakerBoost: asBoolean(talkVoiceSettings?.useSpeakerBoost) }),
|
||||
...(asNumber(talkVoiceSettings?.speed) == null
|
||||
? {}
|
||||
: { speed: asNumber(talkVoiceSettings?.speed) }),
|
||||
},
|
||||
};
|
||||
},
|
||||
resolveTalkOverrides: ({ params }) => {
|
||||
const normalize = trimToUndefined(params.normalize);
|
||||
const language = trimToUndefined(params.language)?.toLowerCase();
|
||||
const voiceSettings = {
|
||||
...(asNumber(params.speed) == null ? {} : { speed: asNumber(params.speed) }),
|
||||
...(asNumber(params.stability) == null ? {} : { stability: asNumber(params.stability) }),
|
||||
...(asNumber(params.similarity) == null
|
||||
? {}
|
||||
: { similarityBoost: asNumber(params.similarity) }),
|
||||
...(asNumber(params.style) == null ? {} : { style: asNumber(params.style) }),
|
||||
...(asBoolean(params.speakerBoost) == null
|
||||
? {}
|
||||
: { useSpeakerBoost: asBoolean(params.speakerBoost) }),
|
||||
};
|
||||
return {
|
||||
...(trimToUndefined(params.voiceId) == null
|
||||
? {}
|
||||
: { voiceId: trimToUndefined(params.voiceId) }),
|
||||
...(trimToUndefined(params.modelId) == null
|
||||
? {}
|
||||
: { modelId: trimToUndefined(params.modelId) }),
|
||||
...(trimToUndefined(params.outputFormat) == null
|
||||
? {}
|
||||
: { outputFormat: trimToUndefined(params.outputFormat) }),
|
||||
...(asNumber(params.seed) == null ? {} : { seed: asNumber(params.seed) }),
|
||||
...(normalize == null
|
||||
? {}
|
||||
: { applyTextNormalization: normalizeApplyTextNormalization(normalize) }),
|
||||
...(language == null ? {} : { languageCode: normalizeLanguageCode(language) }),
|
||||
...(Object.keys(voiceSettings).length === 0 ? {} : { voiceSettings }),
|
||||
};
|
||||
},
|
||||
listVoices: async (req) => {
|
||||
const config = req.providerConfig
|
||||
? readElevenLabsProviderConfig(req.providerConfig)
|
||||
: undefined;
|
||||
const apiKey =
|
||||
req.apiKey ||
|
||||
req.config?.elevenlabs.apiKey ||
|
||||
process.env.ELEVENLABS_API_KEY ||
|
||||
process.env.XI_API_KEY;
|
||||
req.apiKey || config?.apiKey || process.env.ELEVENLABS_API_KEY || process.env.XI_API_KEY;
|
||||
if (!apiKey) {
|
||||
throw new Error("ElevenLabs API key missing");
|
||||
}
|
||||
return listElevenLabsVoices({
|
||||
apiKey,
|
||||
baseUrl: req.baseUrl ?? req.config?.elevenlabs.baseUrl,
|
||||
baseUrl: req.baseUrl ?? config?.baseUrl,
|
||||
});
|
||||
},
|
||||
isConfigured: ({ config }) =>
|
||||
Boolean(config.elevenlabs.apiKey || process.env.ELEVENLABS_API_KEY || process.env.XI_API_KEY),
|
||||
isConfigured: ({ providerConfig }) =>
|
||||
Boolean(
|
||||
readElevenLabsProviderConfig(providerConfig).apiKey ||
|
||||
process.env.ELEVENLABS_API_KEY ||
|
||||
process.env.XI_API_KEY,
|
||||
),
|
||||
synthesize: async (req) => {
|
||||
const apiKey =
|
||||
req.config.elevenlabs.apiKey || process.env.ELEVENLABS_API_KEY || process.env.XI_API_KEY;
|
||||
const config = readElevenLabsProviderConfig(req.providerConfig);
|
||||
const overrides = req.providerOverrides ?? {};
|
||||
const apiKey = config.apiKey || process.env.ELEVENLABS_API_KEY || process.env.XI_API_KEY;
|
||||
if (!apiKey) {
|
||||
throw new Error("ElevenLabs API key missing");
|
||||
}
|
||||
const outputFormat =
|
||||
req.overrides?.elevenlabs?.outputFormat ??
|
||||
trimToUndefined(overrides.outputFormat) ??
|
||||
(req.target === "voice-note" ? "opus_48000_64" : "mp3_44100_128");
|
||||
const overrideVoiceSettings = asObject(overrides.voiceSettings);
|
||||
const audioBuffer = await elevenLabsTTS({
|
||||
text: req.text,
|
||||
apiKey,
|
||||
baseUrl: req.config.elevenlabs.baseUrl,
|
||||
voiceId: req.overrides?.elevenlabs?.voiceId ?? req.config.elevenlabs.voiceId,
|
||||
modelId: req.overrides?.elevenlabs?.modelId ?? req.config.elevenlabs.modelId,
|
||||
baseUrl: config.baseUrl,
|
||||
voiceId: trimToUndefined(overrides.voiceId) ?? config.voiceId,
|
||||
modelId: trimToUndefined(overrides.modelId) ?? config.modelId,
|
||||
outputFormat,
|
||||
seed: req.overrides?.elevenlabs?.seed ?? req.config.elevenlabs.seed,
|
||||
seed: asNumber(overrides.seed) ?? config.seed,
|
||||
applyTextNormalization:
|
||||
req.overrides?.elevenlabs?.applyTextNormalization ??
|
||||
req.config.elevenlabs.applyTextNormalization,
|
||||
languageCode: req.overrides?.elevenlabs?.languageCode ?? req.config.elevenlabs.languageCode,
|
||||
(trimToUndefined(overrides.applyTextNormalization) as
|
||||
| "auto"
|
||||
| "on"
|
||||
| "off"
|
||||
| undefined) ?? config.applyTextNormalization,
|
||||
languageCode: trimToUndefined(overrides.languageCode) ?? config.languageCode,
|
||||
voiceSettings: {
|
||||
...req.config.elevenlabs.voiceSettings,
|
||||
...req.overrides?.elevenlabs?.voiceSettings,
|
||||
...config.voiceSettings,
|
||||
...(asNumber(overrideVoiceSettings?.stability) == null
|
||||
? {}
|
||||
: { stability: asNumber(overrideVoiceSettings?.stability) }),
|
||||
...(asNumber(overrideVoiceSettings?.similarityBoost) == null
|
||||
? {}
|
||||
: { similarityBoost: asNumber(overrideVoiceSettings?.similarityBoost) }),
|
||||
...(asNumber(overrideVoiceSettings?.style) == null
|
||||
? {}
|
||||
: { style: asNumber(overrideVoiceSettings?.style) }),
|
||||
...(asBoolean(overrideVoiceSettings?.useSpeakerBoost) == null
|
||||
? {}
|
||||
: { useSpeakerBoost: asBoolean(overrideVoiceSettings?.useSpeakerBoost) }),
|
||||
...(asNumber(overrideVoiceSettings?.speed) == null
|
||||
? {}
|
||||
: { speed: asNumber(overrideVoiceSettings?.speed) }),
|
||||
},
|
||||
timeoutMs: req.config.timeoutMs,
|
||||
timeoutMs: req.timeoutMs,
|
||||
});
|
||||
return {
|
||||
audioBuffer,
|
||||
@@ -101,8 +509,8 @@ export function buildElevenLabsSpeechProvider(): SpeechProviderPlugin {
|
||||
};
|
||||
},
|
||||
synthesizeTelephony: async (req) => {
|
||||
const apiKey =
|
||||
req.config.elevenlabs.apiKey || process.env.ELEVENLABS_API_KEY || process.env.XI_API_KEY;
|
||||
const config = readElevenLabsProviderConfig(req.providerConfig);
|
||||
const apiKey = config.apiKey || process.env.ELEVENLABS_API_KEY || process.env.XI_API_KEY;
|
||||
if (!apiKey) {
|
||||
throw new Error("ElevenLabs API key missing");
|
||||
}
|
||||
@@ -111,15 +519,15 @@ export function buildElevenLabsSpeechProvider(): SpeechProviderPlugin {
|
||||
const audioBuffer = await elevenLabsTTS({
|
||||
text: req.text,
|
||||
apiKey,
|
||||
baseUrl: req.config.elevenlabs.baseUrl,
|
||||
voiceId: req.config.elevenlabs.voiceId,
|
||||
modelId: req.config.elevenlabs.modelId,
|
||||
baseUrl: config.baseUrl,
|
||||
voiceId: config.voiceId,
|
||||
modelId: config.modelId,
|
||||
outputFormat,
|
||||
seed: req.config.elevenlabs.seed,
|
||||
applyTextNormalization: req.config.elevenlabs.applyTextNormalization,
|
||||
languageCode: req.config.elevenlabs.languageCode,
|
||||
voiceSettings: req.config.elevenlabs.voiceSettings,
|
||||
timeoutMs: req.config.timeoutMs,
|
||||
seed: config.seed,
|
||||
applyTextNormalization: config.applyTextNormalization,
|
||||
languageCode: config.languageCode,
|
||||
voiceSettings: config.voiceSettings,
|
||||
timeoutMs: req.timeoutMs,
|
||||
});
|
||||
return { audioBuffer, outputFormat, sampleRate };
|
||||
},
|
||||
|
||||
@@ -5,14 +5,33 @@ import {
|
||||
TRUSTED_CLIENT_TOKEN,
|
||||
generateSecMsGecToken,
|
||||
} from "node-edge-tts/dist/drm.js";
|
||||
import type { SpeechProviderPlugin } from "openclaw/plugin-sdk/core";
|
||||
import { resolvePreferredOpenClawTmpDir } from "openclaw/plugin-sdk/llm-task";
|
||||
import { isVoiceCompatibleAudio } from "openclaw/plugin-sdk/media-runtime";
|
||||
import type { SpeechVoiceOption } from "openclaw/plugin-sdk/speech";
|
||||
import type {
|
||||
SpeechProviderConfig,
|
||||
SpeechProviderPlugin,
|
||||
SpeechVoiceOption,
|
||||
} from "openclaw/plugin-sdk/speech-core";
|
||||
import { edgeTTS, inferEdgeExtension } from "./tts.js";
|
||||
|
||||
const DEFAULT_EDGE_VOICE = "en-US-MichelleNeural";
|
||||
const DEFAULT_EDGE_LANG = "en-US";
|
||||
const DEFAULT_EDGE_OUTPUT_FORMAT = "audio-24khz-48kbitrate-mono-mp3";
|
||||
|
||||
type MicrosoftProviderConfig = {
|
||||
enabled: boolean;
|
||||
voice: string;
|
||||
lang: string;
|
||||
outputFormat: string;
|
||||
outputFormatConfigured: boolean;
|
||||
pitch?: string;
|
||||
rate?: string;
|
||||
volume?: string;
|
||||
saveSubtitles: boolean;
|
||||
proxy?: string;
|
||||
timeoutMs?: number;
|
||||
};
|
||||
|
||||
type MicrosoftVoiceListEntry = {
|
||||
ShortName?: string;
|
||||
FriendlyName?: string;
|
||||
@@ -24,6 +43,64 @@ type MicrosoftVoiceListEntry = {
|
||||
};
|
||||
};
|
||||
|
||||
function trimToUndefined(value: unknown): string | undefined {
|
||||
return typeof value === "string" && value.trim() ? value.trim() : undefined;
|
||||
}
|
||||
|
||||
function asBoolean(value: unknown): boolean | undefined {
|
||||
return typeof value === "boolean" ? value : undefined;
|
||||
}
|
||||
|
||||
function asNumber(value: unknown): number | undefined {
|
||||
return typeof value === "number" && Number.isFinite(value) ? value : undefined;
|
||||
}
|
||||
|
||||
function asObject(value: unknown): Record<string, unknown> | undefined {
|
||||
return typeof value === "object" && value !== null && !Array.isArray(value)
|
||||
? (value as Record<string, unknown>)
|
||||
: undefined;
|
||||
}
|
||||
|
||||
function normalizeMicrosoftProviderConfig(
|
||||
rawConfig: Record<string, unknown>,
|
||||
): MicrosoftProviderConfig {
|
||||
const rawEdge = asObject(rawConfig.edge);
|
||||
const rawMicrosoft = asObject(rawConfig.microsoft);
|
||||
const raw = { ...(rawEdge ?? {}), ...(rawMicrosoft ?? {}) };
|
||||
const outputFormat = trimToUndefined(raw.outputFormat);
|
||||
return {
|
||||
enabled: asBoolean(raw.enabled) ?? true,
|
||||
voice: trimToUndefined(raw.voice) ?? DEFAULT_EDGE_VOICE,
|
||||
lang: trimToUndefined(raw.lang) ?? DEFAULT_EDGE_LANG,
|
||||
outputFormat: outputFormat ?? DEFAULT_EDGE_OUTPUT_FORMAT,
|
||||
outputFormatConfigured: Boolean(outputFormat),
|
||||
pitch: trimToUndefined(raw.pitch),
|
||||
rate: trimToUndefined(raw.rate),
|
||||
volume: trimToUndefined(raw.volume),
|
||||
saveSubtitles: asBoolean(raw.saveSubtitles) ?? false,
|
||||
proxy: trimToUndefined(raw.proxy),
|
||||
timeoutMs: asNumber(raw.timeoutMs),
|
||||
};
|
||||
}
|
||||
|
||||
function readMicrosoftProviderConfig(config: SpeechProviderConfig): MicrosoftProviderConfig {
|
||||
const defaults = normalizeMicrosoftProviderConfig({});
|
||||
return {
|
||||
enabled: asBoolean(config.enabled) ?? defaults.enabled,
|
||||
voice: trimToUndefined(config.voice) ?? defaults.voice,
|
||||
lang: trimToUndefined(config.lang) ?? defaults.lang,
|
||||
outputFormat: trimToUndefined(config.outputFormat) ?? defaults.outputFormat,
|
||||
outputFormatConfigured:
|
||||
asBoolean(config.outputFormatConfigured) ?? defaults.outputFormatConfigured,
|
||||
pitch: trimToUndefined(config.pitch) ?? defaults.pitch,
|
||||
rate: trimToUndefined(config.rate) ?? defaults.rate,
|
||||
volume: trimToUndefined(config.volume) ?? defaults.volume,
|
||||
saveSubtitles: asBoolean(config.saveSubtitles) ?? defaults.saveSubtitles,
|
||||
proxy: trimToUndefined(config.proxy) ?? defaults.proxy,
|
||||
timeoutMs: asNumber(config.timeoutMs) ?? defaults.timeoutMs,
|
||||
};
|
||||
}
|
||||
|
||||
function buildMicrosoftVoiceHeaders(): Record<string, string> {
|
||||
const major = CHROMIUM_FULL_VERSION.split(".")[0] || "0";
|
||||
return {
|
||||
@@ -77,13 +154,57 @@ export function buildMicrosoftSpeechProvider(): SpeechProviderPlugin {
|
||||
id: "microsoft",
|
||||
label: "Microsoft",
|
||||
aliases: ["edge"],
|
||||
autoSelectOrder: 30,
|
||||
resolveConfig: ({ rawConfig }) => normalizeMicrosoftProviderConfig(rawConfig),
|
||||
resolveTalkConfig: ({ baseTtsConfig, talkProviderConfig }) => {
|
||||
const base = normalizeMicrosoftProviderConfig(baseTtsConfig);
|
||||
return {
|
||||
...base,
|
||||
enabled: true,
|
||||
...(trimToUndefined(talkProviderConfig.voiceId) == null
|
||||
? {}
|
||||
: { voice: trimToUndefined(talkProviderConfig.voiceId) }),
|
||||
...(trimToUndefined(talkProviderConfig.languageCode) == null
|
||||
? {}
|
||||
: { lang: trimToUndefined(talkProviderConfig.languageCode) }),
|
||||
...(trimToUndefined(talkProviderConfig.outputFormat) == null
|
||||
? {}
|
||||
: { outputFormat: trimToUndefined(talkProviderConfig.outputFormat) }),
|
||||
...(trimToUndefined(talkProviderConfig.pitch) == null
|
||||
? {}
|
||||
: { pitch: trimToUndefined(talkProviderConfig.pitch) }),
|
||||
...(trimToUndefined(talkProviderConfig.rate) == null
|
||||
? {}
|
||||
: { rate: trimToUndefined(talkProviderConfig.rate) }),
|
||||
...(trimToUndefined(talkProviderConfig.volume) == null
|
||||
? {}
|
||||
: { volume: trimToUndefined(talkProviderConfig.volume) }),
|
||||
...(trimToUndefined(talkProviderConfig.proxy) == null
|
||||
? {}
|
||||
: { proxy: trimToUndefined(talkProviderConfig.proxy) }),
|
||||
...(asNumber(talkProviderConfig.timeoutMs) == null
|
||||
? {}
|
||||
: { timeoutMs: asNumber(talkProviderConfig.timeoutMs) }),
|
||||
};
|
||||
},
|
||||
resolveTalkOverrides: ({ params }) => ({
|
||||
...(trimToUndefined(params.voiceId) == null
|
||||
? {}
|
||||
: { voice: trimToUndefined(params.voiceId) }),
|
||||
...(trimToUndefined(params.outputFormat) == null
|
||||
? {}
|
||||
: { outputFormat: trimToUndefined(params.outputFormat) }),
|
||||
}),
|
||||
listVoices: async () => await listMicrosoftVoices(),
|
||||
isConfigured: ({ config }) => config.edge.enabled,
|
||||
isConfigured: ({ providerConfig }) => readMicrosoftProviderConfig(providerConfig).enabled,
|
||||
synthesize: async (req) => {
|
||||
const config = readMicrosoftProviderConfig(req.providerConfig);
|
||||
const tempRoot = resolvePreferredOpenClawTmpDir();
|
||||
mkdirSync(tempRoot, { recursive: true, mode: 0o700 });
|
||||
const tempDir = mkdtempSync(path.join(tempRoot, "tts-microsoft-"));
|
||||
let outputFormat = req.overrides?.microsoft?.outputFormat ?? req.config.edge.outputFormat;
|
||||
const overrideVoice = trimToUndefined(req.providerOverrides?.voice);
|
||||
let outputFormat =
|
||||
trimToUndefined(req.providerOverrides?.outputFormat) ?? config.outputFormat;
|
||||
const fallbackOutputFormat =
|
||||
outputFormat !== DEFAULT_EDGE_OUTPUT_FORMAT ? DEFAULT_EDGE_OUTPUT_FORMAT : undefined;
|
||||
|
||||
@@ -95,11 +216,11 @@ export function buildMicrosoftSpeechProvider(): SpeechProviderPlugin {
|
||||
text: req.text,
|
||||
outputPath,
|
||||
config: {
|
||||
...req.config.edge,
|
||||
voice: req.overrides?.microsoft?.voice ?? req.config.edge.voice,
|
||||
...config,
|
||||
voice: overrideVoice ?? config.voice,
|
||||
outputFormat: format,
|
||||
},
|
||||
timeoutMs: req.config.timeoutMs,
|
||||
timeoutMs: req.timeoutMs,
|
||||
});
|
||||
const audioBuffer = readFileSync(outputPath);
|
||||
return {
|
||||
@@ -112,9 +233,9 @@ export function buildMicrosoftSpeechProvider(): SpeechProviderPlugin {
|
||||
|
||||
try {
|
||||
return await runEdge(outputFormat);
|
||||
} catch (err) {
|
||||
} catch (error) {
|
||||
if (!fallbackOutputFormat || fallbackOutputFormat === outputFormat) {
|
||||
throw err;
|
||||
throw error;
|
||||
}
|
||||
outputFormat = fallbackOutputFormat;
|
||||
return await runEdge(outputFormat);
|
||||
|
||||
@@ -137,32 +137,14 @@ function createLiveTtsConfig(): ResolvedTtsConfig {
|
||||
allowNormalization: true,
|
||||
allowSeed: true,
|
||||
},
|
||||
elevenlabs: {
|
||||
baseUrl: "https://api.elevenlabs.io",
|
||||
voiceId: "",
|
||||
modelId: "eleven_multilingual_v2",
|
||||
voiceSettings: {
|
||||
stability: 0.5,
|
||||
similarityBoost: 0.75,
|
||||
style: 0,
|
||||
useSpeakerBoost: true,
|
||||
speed: 1,
|
||||
providerConfigs: {
|
||||
openai: {
|
||||
apiKey: OPENAI_API_KEY,
|
||||
baseUrl: "https://api.openai.com/v1",
|
||||
model: "gpt-4o-mini-tts",
|
||||
voice: "alloy",
|
||||
},
|
||||
},
|
||||
openai: {
|
||||
apiKey: OPENAI_API_KEY,
|
||||
baseUrl: "https://api.openai.com/v1",
|
||||
model: "gpt-4o-mini-tts",
|
||||
voice: "alloy",
|
||||
},
|
||||
edge: {
|
||||
enabled: false,
|
||||
voice: "en-US-AriaNeural",
|
||||
lang: "en-US",
|
||||
outputFormat: "audio-24khz-48kbitrate-mono-mp3",
|
||||
outputFormatConfigured: false,
|
||||
saveSubtitles: false,
|
||||
},
|
||||
maxTextLength: 4_000,
|
||||
timeoutMs: 30_000,
|
||||
};
|
||||
@@ -358,8 +340,9 @@ describeLive("openai plugin live", () => {
|
||||
const audioFile = await speechProvider.synthesize({
|
||||
text: "OpenClaw integration test OK.",
|
||||
cfg,
|
||||
config: ttsConfig,
|
||||
providerConfig: ttsConfig.providerConfigs.openai ?? {},
|
||||
target: "audio-file",
|
||||
timeoutMs: ttsConfig.timeoutMs,
|
||||
});
|
||||
expect(audioFile.outputFormat).toBe("mp3");
|
||||
expect(audioFile.fileExtension).toBe(".mp3");
|
||||
@@ -368,7 +351,8 @@ describeLive("openai plugin live", () => {
|
||||
const telephony = await speechProvider.synthesizeTelephony?.({
|
||||
text: "Telephony check OK.",
|
||||
cfg,
|
||||
config: ttsConfig,
|
||||
providerConfig: ttsConfig.providerConfigs.openai ?? {},
|
||||
timeoutMs: ttsConfig.timeoutMs,
|
||||
});
|
||||
expect(telephony?.outputFormat).toBe("pcm");
|
||||
expect(telephony?.sampleRate).toBe(24_000);
|
||||
@@ -386,8 +370,9 @@ describeLive("openai plugin live", () => {
|
||||
const synthesized = await speechProvider.synthesize({
|
||||
text: "OpenClaw integration test OK.",
|
||||
cfg,
|
||||
config: ttsConfig,
|
||||
providerConfig: ttsConfig.providerConfigs.openai ?? {},
|
||||
target: "audio-file",
|
||||
timeoutMs: ttsConfig.timeoutMs,
|
||||
});
|
||||
|
||||
const transcription = await mediaProvider.transcribeAudio?.({
|
||||
|
||||
@@ -1,16 +1,181 @@
|
||||
import type { SpeechProviderPlugin } from "openclaw/plugin-sdk/core";
|
||||
import { OPENAI_TTS_MODELS, OPENAI_TTS_VOICES, openaiTTS } from "./tts.js";
|
||||
import { normalizeResolvedSecretInputString } from "openclaw/plugin-sdk/secret-input";
|
||||
import type {
|
||||
SpeechDirectiveTokenParseContext,
|
||||
SpeechProviderConfig,
|
||||
SpeechProviderOverrides,
|
||||
SpeechProviderPlugin,
|
||||
} from "openclaw/plugin-sdk/speech-core";
|
||||
import {
|
||||
DEFAULT_OPENAI_BASE_URL,
|
||||
isValidOpenAIModel,
|
||||
isValidOpenAIVoice,
|
||||
normalizeOpenAITtsBaseUrl,
|
||||
OPENAI_TTS_MODELS,
|
||||
OPENAI_TTS_VOICES,
|
||||
openaiTTS,
|
||||
} from "./tts.js";
|
||||
|
||||
type OpenAITtsProviderConfig = {
|
||||
apiKey?: string;
|
||||
baseUrl: string;
|
||||
model: string;
|
||||
voice: string;
|
||||
speed?: number;
|
||||
instructions?: string;
|
||||
};
|
||||
|
||||
type OpenAITtsProviderOverrides = {
|
||||
model?: string;
|
||||
voice?: string;
|
||||
speed?: number;
|
||||
};
|
||||
|
||||
function trimToUndefined(value: unknown): string | undefined {
|
||||
return typeof value === "string" && value.trim() ? value.trim() : undefined;
|
||||
}
|
||||
|
||||
function asNumber(value: unknown): number | undefined {
|
||||
return typeof value === "number" && Number.isFinite(value) ? value : undefined;
|
||||
}
|
||||
|
||||
function asObject(value: unknown): Record<string, unknown> | undefined {
|
||||
return typeof value === "object" && value !== null && !Array.isArray(value)
|
||||
? (value as Record<string, unknown>)
|
||||
: undefined;
|
||||
}
|
||||
|
||||
function normalizeOpenAIProviderConfig(
|
||||
rawConfig: Record<string, unknown>,
|
||||
): OpenAITtsProviderConfig {
|
||||
const raw = asObject(rawConfig.openai);
|
||||
return {
|
||||
apiKey: normalizeResolvedSecretInputString({
|
||||
value: raw?.apiKey,
|
||||
path: "messages.tts.openai.apiKey",
|
||||
}),
|
||||
baseUrl: normalizeOpenAITtsBaseUrl(
|
||||
trimToUndefined(raw?.baseUrl) ??
|
||||
trimToUndefined(process.env.OPENAI_TTS_BASE_URL) ??
|
||||
DEFAULT_OPENAI_BASE_URL,
|
||||
),
|
||||
model: trimToUndefined(raw?.model) ?? "gpt-4o-mini-tts",
|
||||
voice: trimToUndefined(raw?.voice) ?? "coral",
|
||||
speed: asNumber(raw?.speed),
|
||||
instructions: trimToUndefined(raw?.instructions),
|
||||
};
|
||||
}
|
||||
|
||||
function readOpenAIProviderConfig(config: SpeechProviderConfig): OpenAITtsProviderConfig {
|
||||
const normalized = normalizeOpenAIProviderConfig({});
|
||||
return {
|
||||
apiKey: trimToUndefined(config.apiKey) ?? normalized.apiKey,
|
||||
baseUrl: trimToUndefined(config.baseUrl) ?? normalized.baseUrl,
|
||||
model: trimToUndefined(config.model) ?? normalized.model,
|
||||
voice: trimToUndefined(config.voice) ?? normalized.voice,
|
||||
speed: asNumber(config.speed) ?? normalized.speed,
|
||||
instructions: trimToUndefined(config.instructions) ?? normalized.instructions,
|
||||
};
|
||||
}
|
||||
|
||||
function readOpenAIOverrides(
|
||||
overrides: SpeechProviderOverrides | undefined,
|
||||
): OpenAITtsProviderOverrides {
|
||||
if (!overrides) {
|
||||
return {};
|
||||
}
|
||||
return {
|
||||
model: trimToUndefined(overrides.model),
|
||||
voice: trimToUndefined(overrides.voice),
|
||||
speed: asNumber(overrides.speed),
|
||||
};
|
||||
}
|
||||
|
||||
function parseDirectiveToken(ctx: SpeechDirectiveTokenParseContext): {
|
||||
handled: boolean;
|
||||
overrides?: SpeechProviderOverrides;
|
||||
warnings?: string[];
|
||||
} {
|
||||
const baseUrl = trimToUndefined(ctx.providerConfig?.baseUrl);
|
||||
switch (ctx.key) {
|
||||
case "voice":
|
||||
case "openai_voice":
|
||||
case "openaivoice":
|
||||
if (!ctx.policy.allowVoice) {
|
||||
return { handled: true };
|
||||
}
|
||||
if (!isValidOpenAIVoice(ctx.value, baseUrl)) {
|
||||
return { handled: true, warnings: [`invalid OpenAI voice "${ctx.value}"`] };
|
||||
}
|
||||
return { handled: true, overrides: { voice: ctx.value } };
|
||||
case "model":
|
||||
case "openai_model":
|
||||
case "openaimodel":
|
||||
if (!ctx.policy.allowModelId) {
|
||||
return { handled: true };
|
||||
}
|
||||
if (!isValidOpenAIModel(ctx.value, baseUrl)) {
|
||||
return { handled: false };
|
||||
}
|
||||
return { handled: true, overrides: { model: ctx.value } };
|
||||
default:
|
||||
return { handled: false };
|
||||
}
|
||||
}
|
||||
|
||||
export function buildOpenAISpeechProvider(): SpeechProviderPlugin {
|
||||
return {
|
||||
id: "openai",
|
||||
label: "OpenAI",
|
||||
autoSelectOrder: 10,
|
||||
models: OPENAI_TTS_MODELS,
|
||||
voices: OPENAI_TTS_VOICES,
|
||||
resolveConfig: ({ rawConfig }) => normalizeOpenAIProviderConfig(rawConfig),
|
||||
parseDirectiveToken,
|
||||
resolveTalkConfig: ({ baseTtsConfig, talkProviderConfig }) => {
|
||||
const base = normalizeOpenAIProviderConfig(baseTtsConfig);
|
||||
return {
|
||||
...base,
|
||||
...(talkProviderConfig.apiKey === undefined
|
||||
? {}
|
||||
: {
|
||||
apiKey: normalizeResolvedSecretInputString({
|
||||
value: talkProviderConfig.apiKey,
|
||||
path: "talk.providers.openai.apiKey",
|
||||
}),
|
||||
}),
|
||||
...(trimToUndefined(talkProviderConfig.baseUrl) == null
|
||||
? {}
|
||||
: { baseUrl: trimToUndefined(talkProviderConfig.baseUrl) }),
|
||||
...(trimToUndefined(talkProviderConfig.modelId) == null
|
||||
? {}
|
||||
: { model: trimToUndefined(talkProviderConfig.modelId) }),
|
||||
...(trimToUndefined(talkProviderConfig.voiceId) == null
|
||||
? {}
|
||||
: { voice: trimToUndefined(talkProviderConfig.voiceId) }),
|
||||
...(asNumber(talkProviderConfig.speed) == null
|
||||
? {}
|
||||
: { speed: asNumber(talkProviderConfig.speed) }),
|
||||
...(trimToUndefined(talkProviderConfig.instructions) == null
|
||||
? {}
|
||||
: { instructions: trimToUndefined(talkProviderConfig.instructions) }),
|
||||
};
|
||||
},
|
||||
resolveTalkOverrides: ({ params }) => ({
|
||||
...(trimToUndefined(params.voiceId) == null
|
||||
? {}
|
||||
: { voice: trimToUndefined(params.voiceId) }),
|
||||
...(trimToUndefined(params.modelId) == null
|
||||
? {}
|
||||
: { model: trimToUndefined(params.modelId) }),
|
||||
...(asNumber(params.speed) == null ? {} : { speed: asNumber(params.speed) }),
|
||||
}),
|
||||
listVoices: async () => OPENAI_TTS_VOICES.map((voice) => ({ id: voice, name: voice })),
|
||||
isConfigured: ({ config }) => Boolean(config.openai.apiKey || process.env.OPENAI_API_KEY),
|
||||
isConfigured: ({ providerConfig }) =>
|
||||
Boolean(readOpenAIProviderConfig(providerConfig).apiKey || process.env.OPENAI_API_KEY),
|
||||
synthesize: async (req) => {
|
||||
const apiKey = req.config.openai.apiKey || process.env.OPENAI_API_KEY;
|
||||
const config = readOpenAIProviderConfig(req.providerConfig);
|
||||
const overrides = readOpenAIOverrides(req.providerOverrides);
|
||||
const apiKey = config.apiKey || process.env.OPENAI_API_KEY;
|
||||
if (!apiKey) {
|
||||
throw new Error("OpenAI API key missing");
|
||||
}
|
||||
@@ -18,13 +183,13 @@ export function buildOpenAISpeechProvider(): SpeechProviderPlugin {
|
||||
const audioBuffer = await openaiTTS({
|
||||
text: req.text,
|
||||
apiKey,
|
||||
baseUrl: req.config.openai.baseUrl,
|
||||
model: req.overrides?.openai?.model ?? req.config.openai.model,
|
||||
voice: req.overrides?.openai?.voice ?? req.config.openai.voice,
|
||||
speed: req.overrides?.openai?.speed ?? req.config.openai.speed,
|
||||
instructions: req.config.openai.instructions,
|
||||
baseUrl: config.baseUrl,
|
||||
model: overrides.model ?? config.model,
|
||||
voice: overrides.voice ?? config.voice,
|
||||
speed: overrides.speed ?? config.speed,
|
||||
instructions: config.instructions,
|
||||
responseFormat,
|
||||
timeoutMs: req.config.timeoutMs,
|
||||
timeoutMs: req.timeoutMs,
|
||||
});
|
||||
return {
|
||||
audioBuffer,
|
||||
@@ -34,7 +199,8 @@ export function buildOpenAISpeechProvider(): SpeechProviderPlugin {
|
||||
};
|
||||
},
|
||||
synthesizeTelephony: async (req) => {
|
||||
const apiKey = req.config.openai.apiKey || process.env.OPENAI_API_KEY;
|
||||
const config = readOpenAIProviderConfig(req.providerConfig);
|
||||
const apiKey = config.apiKey || process.env.OPENAI_API_KEY;
|
||||
if (!apiKey) {
|
||||
throw new Error("OpenAI API key missing");
|
||||
}
|
||||
@@ -43,13 +209,13 @@ export function buildOpenAISpeechProvider(): SpeechProviderPlugin {
|
||||
const audioBuffer = await openaiTTS({
|
||||
text: req.text,
|
||||
apiKey,
|
||||
baseUrl: req.config.openai.baseUrl,
|
||||
model: req.config.openai.model,
|
||||
voice: req.config.openai.voice,
|
||||
speed: req.config.openai.speed,
|
||||
instructions: req.config.openai.instructions,
|
||||
baseUrl: config.baseUrl,
|
||||
model: config.model,
|
||||
voice: config.voice,
|
||||
speed: config.speed,
|
||||
instructions: config.instructions,
|
||||
responseFormat: outputFormat,
|
||||
timeoutMs: req.config.timeoutMs,
|
||||
timeoutMs: req.timeoutMs,
|
||||
});
|
||||
return { audioBuffer, outputFormat, sampleRate };
|
||||
},
|
||||
|
||||
@@ -1,4 +1,4 @@
|
||||
const DEFAULT_OPENAI_BASE_URL = "https://api.openai.com/v1";
|
||||
export const DEFAULT_OPENAI_BASE_URL = "https://api.openai.com/v1";
|
||||
|
||||
export const OPENAI_TTS_MODELS = ["gpt-4o-mini-tts", "tts-1", "tts-1-hd"] as const;
|
||||
|
||||
@@ -21,7 +21,7 @@ export const OPENAI_TTS_VOICES = [
|
||||
|
||||
type OpenAiTtsVoice = (typeof OPENAI_TTS_VOICES)[number];
|
||||
|
||||
function normalizeOpenAITtsBaseUrl(baseUrl?: string): string {
|
||||
export function normalizeOpenAITtsBaseUrl(baseUrl?: string): string {
|
||||
const trimmed = baseUrl?.trim();
|
||||
if (!trimmed) {
|
||||
return DEFAULT_OPENAI_BASE_URL;
|
||||
@@ -36,21 +36,24 @@ function isCustomOpenAIEndpoint(baseUrl?: string): boolean {
|
||||
return normalizeOpenAITtsBaseUrl(process.env.OPENAI_TTS_BASE_URL) !== DEFAULT_OPENAI_BASE_URL;
|
||||
}
|
||||
|
||||
function isValidOpenAIModel(model: string, baseUrl?: string): boolean {
|
||||
export function isValidOpenAIModel(model: string, baseUrl?: string): boolean {
|
||||
if (isCustomOpenAIEndpoint(baseUrl)) {
|
||||
return true;
|
||||
}
|
||||
return OPENAI_TTS_MODELS.includes(model as (typeof OPENAI_TTS_MODELS)[number]);
|
||||
}
|
||||
|
||||
function isValidOpenAIVoice(voice: string, baseUrl?: string): voice is OpenAiTtsVoice {
|
||||
export function isValidOpenAIVoice(voice: string, baseUrl?: string): voice is OpenAiTtsVoice {
|
||||
if (isCustomOpenAIEndpoint(baseUrl)) {
|
||||
return true;
|
||||
}
|
||||
return OPENAI_TTS_VOICES.includes(voice as OpenAiTtsVoice);
|
||||
}
|
||||
|
||||
function resolveOpenAITtsInstructions(model: string, instructions?: string): string | undefined {
|
||||
export function resolveOpenAITtsInstructions(
|
||||
model: string,
|
||||
instructions?: string,
|
||||
): string | undefined {
|
||||
const next = instructions?.trim();
|
||||
return next && model.includes("gpt-4o-mini-tts") ? next : undefined;
|
||||
}
|
||||
|
||||
@@ -1,4 +1,3 @@
|
||||
import { resolveOpenAITtsInstructions } from "../../api.js";
|
||||
import { convertPcmToMulaw8k } from "../telephony-audio.js";
|
||||
|
||||
/**
|
||||
@@ -72,6 +71,11 @@ function trimToUndefined(value: string | undefined): string | undefined {
|
||||
return trimmed ? trimmed : undefined;
|
||||
}
|
||||
|
||||
function resolveOpenAITtsInstructions(model: string, instructions?: string): string | undefined {
|
||||
const next = trimToUndefined(instructions);
|
||||
return next && model.includes("gpt-4o-mini-tts") ? next : undefined;
|
||||
}
|
||||
|
||||
/**
|
||||
* OpenAI TTS Provider for generating speech audio.
|
||||
*/
|
||||
|
||||
Reference in New Issue
Block a user