feat(tts): add kugelaudio as a TTS vendor

lib/synth-audio.js

@@ -80,8 +80,8 @@ async function synthAudio(client, createHash, retrieveHash, logger, stats, { acc
   logger = logger || noopLogger;
 
   assert.ok(['google', 'aws', 'polly', 'microsoft', 'wellsaid', 'nvidia', 'elevenlabs',
-    'whisper', 'deepgram', 'deepgramflux', 'rimelabs', 'cartesia', 'gradium', 'nineninesix', 'inworld', 'resemble',
-    'murf', 'xai', 'fishaudio']
+    'whisper', 'deepgram', 'deepgramflux', 'rimelabs', 'cartesia', 'gradium', 'kugelaudio', 'nineninesix', 'inworld',
+    'resemble', 'murf', 'xai', 'fishaudio']
     .includes(vendor) ||
   vendor.startsWith('custom'),
   `synthAudio supported vendors are google, aws, microsoft, nvidia and wellsaid ..etc, not ${vendor}`);
@@ -139,6 +139,9 @@ async function synthAudio(client, createHash, retrieveHash, logger, stats, { acc
   } else if ('gradium' === vendor) {
     assert.ok(voice, 'synthAudio requires voice when gradium is used');
     assert.ok(credentials.api_key, 'synthAudio requires api_key when gradium is used');
+  } else if ('kugelaudio' === vendor) {
+    assert.ok(voice, 'synthAudio requires voice when kugelaudio is used');
+    assert.ok(credentials.api_key, 'synthAudio requires api_key when kugelaudio is used');
   } else if ('nineninesix' === vendor) {
     assert.ok(voice, 'synthAudio requires voice when nineninesix is used');
     assert.ok(credentials.api_key, 'synthAudio requires api_key when nineninesix is used');
@@ -234,6 +237,11 @@ async function synthAudio(client, createHash, retrieveHash, logger, stats, { acc
           credentials, options, stats, voice, key, text, renderForCaching, disableTtsStreaming,
           disableTtsCache});
         break;
+      case 'kugelaudio':
+        audioData = await synthKugelaudio(logger, {
+          credentials, options, stats, language, voice, key, text, renderForCaching, disableTtsStreaming,
+          disableTtsCache});
+        break;
       case 'nineninesix':
         audioData = await synthNineninesix(logger, {
           credentials, stats, language, voice, key, text, renderForCaching, disableTtsStreaming,
@@ -1512,6 +1520,88 @@ const synthGradium = async(logger, {
   }
 };
 
+/* kugelaudio — json websocket (/ws/tts/stream) for streaming, and POST /v1/tts/generate
+   for the cache render. the POST streams back bare little-endian 16-bit samples at the
+   requested sample_rate, which is exactly the r8 container at 8000.
+
+   voices are numeric ids (or public handles). language is an ISO 639-1 code that drives
+   text normalization; jambonz carries BCP-47, so only the primary subtag is sent. the
+   api rejects codes outside its list, so an unsupported or unset language is omitted
+   and the voice's own language applies. options.api_uri pins a region
+   (e.g. api.eu.kugelaudio.com).
+*/
+const KUGELAUDIO_LANGUAGES = ['ar', 'bg', 'bn', 'cs', 'da', 'de', 'el', 'en', 'es', 'fa', 'fi', 'fr', 'he', 'hi',
+  'hr', 'hu', 'id', 'it', 'ja', 'ko', 'ms', 'nl', 'no', 'pl', 'pt', 'ro', 'ru', 'sk', 'sl', 'sr', 'sv', 'ta', 'th',
+  'tr', 'uk', 'ur', 'vi', 'yue', 'zh'];
+const synthKugelaudio = async(logger, {
+  credentials, options, stats, language, voice, key, text, renderForCaching, disableTtsStreaming,
+  disableTtsCache
+}) => {
+  const {api_key, model_id} = credentials;
+  const {api_uri, speed, cfg_scale, temperature, normalize, project_id, dictionary_ids} = options || {};
+  const isSet = (v) => v !== null && v !== undefined;
+
+  /* default to using the streaming interface, unless disabled by env var OR we want just a cache file */
+  if (!JAMBONES_DISABLE_TTS_STREAMING && !renderForCaching && !disableTtsStreaming) {
+    let params = '{';
+    params += `api_key=${api_key}`;
+    params += `,playback_id=${key}`;
+    params += ',vendor=kugelaudio';
+    params += `,voice=${voice}`;
+    params += `,write_cache_file=${disableTtsCache ? 0 : 1}`;
+    params += `,model_id=${model_id || 'kugel-3'}`;
+    if (language) params += `,language=${language}`;
+    if (api_uri) params += `,api_uri=${api_uri}`;
+    if (isSet(speed)) params += `,speed=${speed}`;
+    if (isSet(cfg_scale)) params += `,cfg_scale=${cfg_scale}`;
+    if (isSet(temperature)) params += `,temperature=${temperature}`;
+    if (isSet(normalize)) params += `,normalize=${normalize}`;
+    if (isSet(project_id)) params += `,project_id=${project_id}`;
+    /* the say: param parser is bracket-aware, so the json array survives intact */
+    if (Array.isArray(dictionary_ids)) params += `,dictionary_ids=${JSON.stringify(dictionary_ids)}`;
+    params += '}';
+
+    return {
+      filePath: `say:${params}${text.replace(/\n/g, ' ').replace(/\r/g, ' ')}`,
+      servedFromCache: false,
+      rtt: 0
+    };
+  }
+
+  try {
+    const sampleRate = 8000;
+    const host = (api_uri || 'api.kugelaudio.com').replace(/^[a-z]+:\/\//, '').replace(/\/$/, '');
+    const post = bent(`https://${host}`, 'POST', 'buffer', {
+      'Authorization': `Bearer ${api_key}`,
+      'Content-Type': 'application/json; charset=utf-8'
+    });
+    const voiceId = /^\d+$/.test(`${voice}`) ? Number(voice) : voice;
+    const lang = language && language.split('-')[0].toLowerCase();
+    const audioContent = await post('/v1/tts/generate', {
+      text,
+      voice_id: voiceId,
+      model_id: model_id || 'kugel-3',
+      sample_rate: sampleRate,
+      ...(KUGELAUDIO_LANGUAGES.includes(lang) && {language: lang}),
+      ...(isSet(speed) && {speed: Number(speed)}),
+      ...(isSet(cfg_scale) && {cfg_scale: Number(cfg_scale)}),
+      ...(isSet(temperature) && {temperature: Number(temperature)}),
+      ...(isSet(normalize) && {normalize: normalize === true || normalize === 'true'}),
+      ...(isSet(project_id) && {project_id: Number(project_id)}),
+      ...(Array.isArray(dictionary_ids) && {dictionary_ids})
+    });
+    return {
+      audioContent,
+      extension: 'r8',
+      sampleRate
+    };
+  } catch (err) {
+    logger.info({err}, 'synth kugelaudio returned error');
+    stats.increment('tts.count', ['vendor:kugelaudio', 'accepted:no']);
+    throw err;
+  }
+};
+
 /* nineninesix.ai — a Cartesia-compatible API, but only raw/wav come back
    (mp3 is rejected), so the cache render asks for wav rather than mp3. */
 /* fish.audio — msgpack websocket for streaming, and a POST endpoint for the cache

test/synth.js

@@ -1090,6 +1090,36 @@ test('gradium speech synth tests', async(t) => {
   client.quit();
 });
 
+test('kugelaudio speech synth tests', async(t) => {
+  const fn = require('..');
+  const {synthAudio, client} = fn(opts, logger);
+
+  if (!process.env.KUGELAUDIO_API_KEY) {
+    t.pass('skipping kugelaudio speech synth tests since KUGELAUDIO_API_KEY is not provided');
+    return t.end();
+  }
+  const text = 'Guten Tag und willkommen bei jambonz! Ihre Bestellung kostet 12,99 Euro. ' + Date.now();
+  try {
+    const opts = await synthAudio(stats, {
+      vendor: 'kugelaudio',
+      credentials: {
+        api_key: process.env.KUGELAUDIO_API_KEY,
+        model_id: 'kugel-3'
+      },
+      language: 'de-DE',
+      voice: '1930',
+      text,
+      renderForCaching: true
+    });
+    t.ok(!opts.servedFromCache, `successfully synthed kugelaudio audio to ${opts.filePath}`);
+
+  } catch (err) {
+    console.error(JSON.stringify(err));
+    t.end(err);
+  }
+  client.quit();
+});
+
 test('fishaudio speech synth tests', async(t) => {
   const fn = require('..');
   const {synthAudio, client} = fn(opts, logger);