0.2.1

remove audio extension from audio key

lib/add-file-to-cache.js

@@ -3,6 +3,27 @@ const {noopLogger, makeSynthKey} = require('./utils');
 const {JAMBONES_TTS_CACHE_DURATION_MINS} = require('./config');
 const EXPIRES = JAMBONES_TTS_CACHE_DURATION_MINS;
 
+function getExtensionAndSampleRate(path) {
+  const match = path.match(/\.([^.]*)$/);
+  if (!match) {
+    //default should be wav file.
+    return ['wav', 8000];
+  }
+
+  const extension = match[1];
+  const sampleRateMap = {
+    r8: 8000,
+    r16: 16000,
+    r24: 24000,
+    r44: 44100,
+    r48: 48000,
+    r96: 96000,
+  };
+
+  const sampleRate = sampleRateMap[extension] || 8000;
+  return [extension, sampleRate];
+}
+
 async function addFileToCache(client, logger, path,
   {account_sid, vendor, language, voice, deploymentId, engine, text}) {
   let key;
@@ -17,8 +38,15 @@ async function addFileToCache(client, logger, path,
       engine,
       text,
     });
+    const [extension, sampleRate] = getExtensionAndSampleRate(path);
     const audioBuffer = await fs.readFile(path);
-    await client.setex(key, EXPIRES, audioBuffer.toString('base64'));
+    await client.setex(key, EXPIRES, JSON.stringify(
+      {
+        audioContent: audioBuffer.toString('base64'),
+        extension,
+        sampleRate
+      }
+    ));
   } catch (err) {
     logger.error(err, 'addFileToCache: Error');
     return;

lib/get-aws-sts-token.js

@@ -7,22 +7,23 @@ const CACHE_EXPIRY = process.env.AWS_STS_SESSION_RESET_EXPIRY || (EXPIRY - 600);
 
 async function getAwsAuthToken(
   logger, createHash, retrieveHash,
-  {accessKeyId, secretAccessKey, region, roleArn}) {
+  {speech_credential_sid, accessKeyId, secretAccessKey, region, roleArn}) {
   logger = logger || noopLogger;
   try {
-    const key = makeAwsKey(roleArn || accessKeyId);
+    // if incase instance profile is used, speech_credential_sid will be used as key to lookup cache
+    const key = makeAwsKey(roleArn || accessKeyId || speech_credential_sid);
     const obj = await retrieveHash(key);
     if (obj) return {...obj, servedFromCache: true};
-
+    /* access token not found in cache, so generate it using STS */
     let data;
+    let expiry = CACHE_EXPIRY;
     if (roleArn) {
       const stsClient = new STSClient({ region });
       const roleToAssume = { RoleArn: roleArn, RoleSessionName: 'Jambonz_Speech', DurationSeconds: EXPIRY};
       const command = new AssumeRoleCommand(roleToAssume);
 
       data = await stsClient.send(command);
-    } else {
-      /* access token not found in cache, so generate it using STS */
+    } else if (accessKeyId) {
       const stsClient = new STSClient({
         region,
         credentials: {
@@ -32,6 +33,26 @@ async function getAwsAuthToken(
       });
       const command = new GetSessionTokenCommand({DurationSeconds: EXPIRY});
       data = await stsClient.send(command);
+    } else {
+      // instance profile is used.
+      const stsClient = new STSClient({ region });
+      const cred = await stsClient.config.credentials();
+      // method in the AWS SDK automatically fetches credentials using the default credential
+      // provider chain. If the credentials come from an instance profile or an environment
+      // variable, their expiration is controlled by AWS and not explicitly by our code.
+      if (cred && cred.expiration) {
+        const currentTime = new Date();
+        const expiryTime = new Date(cred.expiration);
+        const remainingTimeInSeconds = Math.round((expiryTime - currentTime) / 1000);
+        expiry = remainingTimeInSeconds;
+      }
+      data = {
+        Credentials: {
+          AccessKeyId: cred.accessKeyId,
+          SecretAccessKey: cred.secretAccessKey,
+          SessionToken: cred.sessionToken
+        }
+      };
     }
 
     const credentials = {
@@ -40,9 +61,11 @@ async function getAwsAuthToken(
       sessionToken: data.Credentials.SessionToken,
       securityToken: data.Credentials.SessionToken
     };
-
-    createHash(key, credentials, CACHE_EXPIRY)
-      .catch((err) => logger.error(err, `Error saving hash for key ${key}`));
+    // Only cache if expiry is good
+    if (expiry > 0) {
+      createHash(key, credentials, expiry)
+        .catch((err) => logger.error(err, `Error saving hash for key ${key}`));
+    }
 
     return {...credentials, servedFromCache: false};
   } catch (err) {

lib/synth-audio.js

@@ -3,6 +3,7 @@ const fs = require('fs');
 const bent = require('bent');
 const ttsGoogle = require('@google-cloud/text-to-speech');
 const { PollyClient, SynthesizeSpeechCommand } = require('@aws-sdk/client-polly');
+const { CartesiaClient } = require('@cartesia/cartesia-js');
 
 const sdk = require('microsoft-cognitiveservices-speech-sdk');
 const TextToSpeechV1 = require('ibm-watson/text-to-speech/v1');
@@ -45,7 +46,7 @@ const {
   JAMBONES_HTTP_PROXY_IP,
   JAMBONES_HTTP_PROXY_PORT,
   JAMBONES_TTS_CACHE_DURATION_MINS,
-  JAMBONES_EAGERLY_PRE_CACHE_AUDIO,
+  JAMBONES_TTS_TRIM_SILENCE,
 } = require('./config');
 const EXPIRES = JAMBONES_TTS_CACHE_DURATION_MINS;
 const OpenAI = require('openai');
@@ -90,13 +91,13 @@ async function synthAudio(client, createHash, retrieveHash, logger, stats, { acc
   vendor, language, voice, gender, text, engine, salt, model, credentials, deploymentId,
   disableTtsCache, renderForCaching = false, disableTtsStreaming, options
 }) {
-  let audioBuffer;
+  let audioData;
   let servedFromCache = false;
   let rtt;
   logger = logger || noopLogger;
 
   assert.ok(['google', 'aws', 'polly', 'microsoft', 'wellsaid', 'nuance', 'nvidia', 'ibm', 'elevenlabs',
-    'whisper', 'deepgram', 'playht', 'rimelabs', 'verbio'].includes(vendor) ||
+    'whisper', 'deepgram', 'playht', 'rimelabs', 'verbio', 'cartesia'].includes(vendor) ||
   vendor.startsWith('custom'),
   `synthAudio supported vendors are google, aws, microsoft, nuance, nvidia and wellsaid ..etc, not ${vendor}`);
   if ('google' === vendor) {
@@ -157,7 +158,9 @@ async function synthAudio(client, createHash, retrieveHash, logger, stats, { acc
     if (!credentials.deepgram_tts_uri) {
       assert.ok(credentials.api_key, 'synthAudio requires api_key when deepgram is used');
     }
-
+  } else if ('cartesia' === vendor) {
+    assert.ok(credentials.api_key, 'synthAudio requires api_key when cartesia is used');
+    assert.ok(credentials.model_id, 'synthAudio requires model_id when cartesia is used');
   }
 
   const key = makeSynthKey({
@@ -166,45 +169,24 @@ async function synthAudio(client, createHash, retrieveHash, logger, stats, { acc
     language: language || '',
     voice: voice || deploymentId,
     engine,
-    text,
-    renderForCaching
+    text
   });
-  let filePath;
-  filePath = makeFilePath({vendor, key, salt, renderForCaching});
+
   debug(`synth key is ${key}`);
   let cached;
   if (!disableTtsCache) {
     cached = await client.get(key);
-    /**
-    *  If we are using tts streaming and also precaching audio, audio could have been cached by streaming (r8)
-    * or here in speech-utils due to precaching (mp3), so we need to check for both keys.
-    */
-    if (!cached && JAMBONES_EAGERLY_PRE_CACHE_AUDIO) {
-      const preCachekey = makeSynthKey({
-        account_sid,
-        vendor,
-        language: language || '',
-        voice: voice || deploymentId,
-        engine,
-        text,
-        renderForCaching: true
-      });
-      cached = await client.get(preCachekey);
-      if (cached) {
-        // Precache audio is available update filpath with precache file extension.
-        filePath = makeFilePath({vendor, key, salt, renderForCaching: true});
-      }
-    }
   }
   if (cached) {
     // found in cache - extend the expiry and use it
     debug('result WAS found in cache');
     servedFromCache = true;
     stats.increment('tts.cache.requests', ['found:yes']);
-    audioBuffer = Buffer.from(cached, 'base64');
+    audioData = JSON.parse(cached);
+    // convert base64 audio to buffer
+    audioData.audioContent = Buffer.from(audioData.audioContent, 'base64');
     client.expire(key, EXPIRES).catch((err) => logger.info(err, 'Error setting expires'));
-  }
-  if (!cached) {
+  } else {
     // not found in cache - go get it from speech vendor and add to cache
     debug('result was NOT found in cache');
     stats.increment('tts.cache.requests', ['found:no']);
@@ -212,82 +194,92 @@ async function synthAudio(client, createHash, retrieveHash, logger, stats, { acc
     const startAt = process.hrtime();
     switch (vendor) {
       case 'google':
-        audioBuffer = await synthGoogle(logger, {credentials, stats, language, voice, gender, text});
+        audioData = await synthGoogle(logger, {credentials, stats, language, voice, gender, text});
         break;
       case 'aws':
       case 'polly':
         vendorLabel = 'aws';
-        audioBuffer = await synthPolly(createHash, retrieveHash, logger,
+        audioData = await synthPolly(createHash, retrieveHash, logger,
           {credentials, stats, language, voice, text, engine});
         break;
       case 'azure':
       case 'microsoft':
         vendorLabel = 'microsoft';
-        audioBuffer = await synthMicrosoft(logger, {credentials, stats, language, voice, text, deploymentId,
-          filePath, renderForCaching, disableTtsStreaming});
+        audioData = await synthMicrosoft(logger, {credentials, stats, language, voice, text, deploymentId,
+          renderForCaching, disableTtsStreaming});
         break;
       case 'nuance':
         model = model || 'enhanced';
-        audioBuffer = await synthNuance(client, logger, {credentials, stats, voice, model, text});
+        audioData = await synthNuance(client, logger, {credentials, stats, voice, model, text});
         break;
       case 'nvidia':
-        audioBuffer = await synthNvidia(client, logger, {credentials, stats, language, voice, model, text});
+        audioData = await synthNvidia(client, logger, {credentials, stats, language, voice, model, text});
         break;
       case 'ibm':
-        audioBuffer = await synthIbm(logger, {credentials, stats, voice, text});
+        audioData = await synthIbm(logger, {credentials, stats, voice, text});
         break;
       case 'wellsaid':
-        audioBuffer = await synthWellSaid(logger, {credentials, stats, language, voice, text, filePath});
+        audioData = await synthWellSaid(logger, {credentials, stats, language, voice, text});
         break;
       case 'elevenlabs':
-        audioBuffer = await synthElevenlabs(logger, {
-          credentials, options, stats, language, voice, text, renderForCaching, disableTtsStreaming, filePath
-        });
+        audioData = await synthElevenlabs(logger, {
+          credentials, options, stats, language, voice, text, renderForCaching, disableTtsStreaming});
         break;
       case 'playht':
-        audioBuffer = await synthPlayHT(client, logger, {
-          credentials, options, stats, language, voice, text, renderForCaching, disableTtsStreaming, filePath
-        });
+        audioData = await synthPlayHT(client, logger, {
+          credentials, options, stats, language, voice, text, renderForCaching, disableTtsStreaming});
+        break;
+      case 'cartesia':
+        audioData = await synthCartesia(logger, {
+          credentials, options, stats, language, voice, text, renderForCaching, disableTtsStreaming});
         break;
       case 'rimelabs':
-        audioBuffer = await synthRimelabs(logger, {
-          credentials, options, stats, language, voice, text, renderForCaching, disableTtsStreaming, filePath
-        });
+        audioData = await synthRimelabs(logger, {
+          credentials, options, stats, language, voice, text, renderForCaching, disableTtsStreaming});
         break;
       case 'whisper':
-        audioBuffer = await synthWhisper(logger, {
+        audioData = await synthWhisper(logger, {
           credentials, stats, voice, text, renderForCaching, disableTtsStreaming});
         break;
       case 'verbio':
-        audioBuffer = await synthVerbio(client, logger, {
+        audioData = await synthVerbio(client, logger, {
           credentials, stats, voice, text, renderForCaching, disableTtsStreaming});
-        if (audioBuffer?.filePath) return audioBuffer;
+        if (audioData?.filePath) return audioData;
         break;
       case 'deepgram':
-        audioBuffer = await synthDeepgram(logger, {credentials, stats, model, text,
+        audioData = await synthDeepgram(logger, {credentials, stats, model, text,
           renderForCaching, disableTtsStreaming});
         break;
       case vendor.startsWith('custom') ? vendor : 'cant_match_value':
-        ({ audioBuffer, filePath } = await synthCustomVendor(logger,
-          {credentials, stats, language, voice, text, filePath}));
+        audioData = await synthCustomVendor(logger,
+          {credentials, stats, language, voice, text});
         break;
       default:
         assert(`synthAudio: unsupported speech vendor ${vendor}`);
     }
-    if ('filePath' in audioBuffer) return audioBuffer;
+    if ('filePath' in audioData) return audioData;
     const diff = process.hrtime(startAt);
     const time = diff[0] * 1e3 + diff[1] * 1e-6;
     rtt = time.toFixed(0);
     stats.histogram('tts.response_time', rtt, [`vendor:${vendorLabel}`]);
     debug(`tts rtt time for ${text.length} chars on ${vendorLabel}: ${rtt}`);
     logger.info(`tts rtt time for ${text.length} chars on ${vendorLabel}: ${rtt}`);
-
-    client.setex(key, EXPIRES, audioBuffer.toString('base64'))
+    // Save audio json to cache
+    client.setex(key, EXPIRES, JSON.stringify({
+      ...audioData,
+      audioContent: audioData.audioContent?.toString('base64')
+    }))
       .catch((err) => logger.error(err, `error calling setex on key ${key}`));
   }
 
   return new Promise((resolve, reject) => {
-    fs.writeFile(filePath, audioBuffer, (err) => {
+    const { audioContent, extension } = audioData;
+    const filePath = makeFilePath({
+      key,
+      salt,
+      extension
+    });
+    fs.writeFile(filePath, audioContent, (err) => {
       if (err) return reject(err);
       resolve({filePath, servedFromCache, rtt});
     });
@@ -342,7 +334,13 @@ const synthPolly = async(createHash, retrieveHash, logger,
         .on('data', (chunk) => {
           chunks.push(chunk);
         })
-        .on('end', () => resolve(Buffer.concat(chunks)));
+        .on('end', () => resolve(
+          {
+            audioContent: Buffer.concat(chunks),
+            extension: 'mp3',
+            sampleRate: 8000
+          }
+        ));
     });
   } catch (err) {
     logger.info({err}, 'synthAudio: Error synthesizing speech using aws polly');
@@ -353,6 +351,48 @@ const synthPolly = async(createHash, retrieveHash, logger,
 
 const synthGoogle = async(logger, {credentials, stats, language, voice, gender, text}) => {
   const client = new ttsGoogle.TextToSpeechClient(credentials);
+  // If google custom voice cloning is used.
+  // At this time 31 Oct 2024, google node sdk has not support voice cloning yet.
+  if (typeof voice === 'object' && voice.voice_cloning_key) {
+    try {
+      const accessToken = await client.auth.getAccessToken();
+      const projectId = await client.getProjectId();
+
+      const post = bent('https://texttospeech.googleapis.com', 'POST', 'json', {
+        'Authorization': `Bearer ${accessToken}`,
+        'x-goog-user-project': projectId,
+        'Content-Type': 'application/json; charset=utf-8'
+      });
+
+      const payload = {
+        input: {
+          text
+        },
+        voice: {
+          language_code: language,
+          voice_clone: {
+            voice_cloning_key: voice.voice_cloning_key
+          }
+        },
+        audioConfig: {
+          // Cloning voice at this time still in v1 beta version, and it support LINEAR16 in Wav format, 24.000Hz
+          audioEncoding: 'LINEAR16',
+          sample_rate_hertz: 24000
+        }
+      };
+
+      const wav = await post('/v1beta1/text:synthesize', payload);
+      return {
+        audioContent: Buffer.from(wav.audioContent, 'base64'),
+        extension: 'wav',
+        sampleRate: 24000
+      };
+    } catch (err) {
+      logger.info({err: await err.text()}, 'synthGoogle returned error');
+      throw err;
+    }
+  }
+
   const opts = {
     voice: {
       ...(typeof voice === 'string' && {name: voice}),
@@ -367,7 +407,11 @@ const synthGoogle = async(logger, {credentials, stats, language, voice, gender,
     const responses = await client.synthesizeSpeech(opts);
     stats.increment('tts.count', ['vendor:google', 'accepted:yes']);
     client.close();
-    return responses[0].audioContent;
+    return {
+      audioContent: responses[0].audioContent,
+      extension: 'mp3',
+      sampleRate: 8000
+    };
   } catch (err) {
     console.error(err);
     logger.info({err, opts}, 'synthAudio: Error synthesizing speech using google');
@@ -398,7 +442,11 @@ const synthIbm = async(logger, {credentials, stats, voice, text}) => {
     for await (const chunk of r.result) {
       chunks.push(chunk);
     }
-    return Buffer.concat(chunks);
+    return {
+      audioContent: Buffer.concat(chunks),
+      extension: 'mp3',
+      sampleRate: 8000
+    };
   } catch (err) {
     logger.info({err, params}, 'synthAudio: Error synthesizing speech using ibm');
     stats.increment('tts.count', ['vendor:ibm', 'accepted:no']);
@@ -408,11 +456,9 @@ const synthIbm = async(logger, {credentials, stats, voice, text}) => {
 
 async function _synthOnPremMicrosoft(logger, {
   credentials,
-  stats,
   language,
   voice,
-  text,
-  filePath
+  text
 }) {
   const {use_custom_tts, custom_tts_endpoint_url, api_key} = credentials;
   let content = text;
@@ -434,15 +480,19 @@ async function _synthOnPremMicrosoft(logger, {
   }
 
   try {
-    const trimSilence = filePath.endsWith('.r8');
+    const trimSilence = JAMBONES_TTS_TRIM_SILENCE;
     const post = bent('POST', 'buffer', {
       'X-Microsoft-OutputFormat': trimSilence ? 'raw-8khz-16bit-mono-pcm' : 'audio-16khz-32kbitrate-mono-mp3',
       'Content-Type': 'application/ssml+xml',
       'User-Agent': 'Jambonz',
       ...(api_key && {'Ocp-Apim-Subscription-Key': api_key})
     });
-    const mp3 = await post(custom_tts_endpoint_url, content);
-    return mp3;
+    const audioContent = await post(custom_tts_endpoint_url, content);
+    return {
+      audioContent,
+      extension: trimSilence ? 'r8' : 'mp3',
+      sampleRate: 8000
+    };
   } catch (err) {
     logger.info({err}, '_synthMicrosoftByHttp returned error');
     throw err;
@@ -455,7 +505,6 @@ const synthMicrosoft = async(logger, {
   language,
   voice,
   text,
-  filePath,
   renderForCaching,
   disableTtsStreaming
 }) => {
@@ -498,17 +547,18 @@ const synthMicrosoft = async(logger, {
         rtt: 0
       };
     }
+    // Azure Onprem
     if (use_custom_tts && custom_tts_endpoint_url) {
       return await _synthOnPremMicrosoft(logger, {
         credentials,
         stats,
         language,
         voice,
-        text,
-        filePath
+        text
       });
     }
-    const trimSilence = filePath.endsWith('.r8');
+    // Azure hosted service
+    const trimSilence = JAMBONES_TTS_TRIM_SILENCE;
     const speechConfig = SpeechConfig.fromSubscription(apiKey, region);
     speechConfig.speechSynthesisLanguage = language;
     speechConfig.speechSynthesisVoiceName = voice;
@@ -543,7 +593,11 @@ const synthMicrosoft = async(logger, {
             case ResultReason.SynthesizingAudioCompleted:
               let buffer = Buffer.from(result.audioData);
               if (trimSilence) buffer = trimTrailingSilence(buffer);
-              resolve(buffer);
+              resolve({
+                audioContent: buffer,
+                extension: trimSilence ? 'r8' : 'mp3',
+                sampleRate: 8000
+              });
               synthesizer.close();
               stats.increment('tts.count', ['vendor:microsoft', 'accepted:yes']);
               break;
@@ -573,11 +627,15 @@ const synthWellSaid = async(logger, {credentials, stats, language, voice, gender
       'Accept': 'audio/mpeg',
       'Content-Type': 'application/json'
     });
-    const mp3 = await post('/v1/tts/stream', {
+    const audioContent = await post('/v1/tts/stream', {
       text,
       speaker_id: voice
     });
-    return mp3;
+    return {
+      audioContent,
+      extension: 'mp3',
+      sampleRate: 8000
+    };
   } catch (err) {
     logger.info({err}, 'testWellSaidTts returned error');
     throw err;
@@ -614,8 +672,8 @@ const synthNuance = async(client, logger, {credentials, stats, voice, model, tex
     t.setText(text);
     input.setText(t);
   }
-
-  pcm.setSampleRateHz(8000);
+  const sampleRate = 8000;
+  pcm.setSampleRateHz(sampleRate);
   f.setPcm(pcm);
   p.setAudioFormat(f);
   v.setName(voice);
@@ -639,7 +697,11 @@ const synthNuance = async(client, logger, {credentials, stats, voice, model, tex
         const details = status.getDetails();
         return reject({code, message, details});
       }
-      resolve(Buffer.from(response.getAudio()));
+      resolve({
+        audioContent: Buffer.from(response.getAudio()),
+        extension: 'r8',
+        sampleRate
+      });
     });
   });
 };
@@ -647,12 +709,13 @@ const synthNuance = async(client, logger, {credentials, stats, voice, model, tex
 const synthNvidia = async(client, logger, {credentials, stats, language,  voice, model, text}) => {
   const {riva_server_uri} = credentials;
   let rivaClient, request;
+  const sampleRate = 8000;
   try {
     rivaClient = await createRivaClient(riva_server_uri);
     request = new SynthesizeSpeechRequest();
     request.setVoiceName(voice);
     request.setLanguageCode(language);
-    request.setSampleRateHz(8000);
+    request.setSampleRateHz(sampleRate);
     request.setEncoding(AudioEncoding.LINEAR_PCM);
     request.setText(text);
   } catch (err) {
@@ -666,7 +729,11 @@ const synthNvidia = async(client, logger, {credentials, stats, language,  voice,
         logger.info({err, voice, language}, 'error synthesizing speech using Nvidia');
         return reject(err);
       }
-      resolve(Buffer.from(response.getAudio()));
+      resolve({
+        audioContent: Buffer.from(response.getAudio()),
+        extension: 'r8',
+        sampleRate
+      });
     });
   });
 };
@@ -688,12 +755,13 @@ const synthCustomVendor = async(logger, {credentials, stats, language, voice, te
       text
     });
 
-    const regex = /\.[^\.]*$/g;
     const mime = response.headers['content-type'];
     const buffer = await response.arrayBuffer();
+    const [extension, sampleRate] = getFileExtFromMime(mime);
     return {
-      audioBuffer: buffer,
-      filePath: filePath.replace(regex, getFileExtFromMime(mime))
+      audioContent: buffer,
+      extension,
+      sampleRate
     };
   } catch (err) {
     logger.info({err}, `Vendor ${vendor} returned error`);
@@ -739,7 +807,7 @@ const synthElevenlabs = async(logger, {
       'Accept': 'audio/mpeg',
       'Content-Type': 'application/json'
     });
-    const mp3 = await post(`/v1/text-to-speech/${voice}${optimize_streaming_latency}`, {
+    const audioContent = await post(`/v1/text-to-speech/${voice}${optimize_streaming_latency}`, {
       text,
       model_id,
       voice_settings: {
@@ -748,7 +816,11 @@ const synthElevenlabs = async(logger, {
       },
       ...opts
     });
-    return mp3;
+    return {
+      audioContent,
+      extension: 'mp3',
+      sampleRate: 8000
+    };
   } catch (err) {
     logger.info({err}, 'synth Elevenlabs returned error');
     stats.increment('tts.count', ['vendor:elevenlabs', 'accepted:no']);
@@ -830,7 +902,7 @@ const synthPlayHT = async(client, logger, {
       'Content-Type': 'application/json'
     });
 
-    const mp3 = await post(synthesizeUrl, {
+    const audioContent = await post(synthesizeUrl, {
       text,
       ...(voice_engine === 'Play3.0' && { language }),
       voice,
@@ -839,7 +911,11 @@ const synthPlayHT = async(client, logger, {
       sample_rate: 8000,
       ...opts
     });
-    return mp3;
+    return {
+      audioContent,
+      extension: 'mp3',
+      sampleRate: 8000
+    };
   } catch (err) {
     logger.info({err}, 'synth PlayHT returned error');
     stats.increment('tts.count', ['vendor:playht', 'accepted:no']);
@@ -878,14 +954,19 @@ const synthRimelabs = async(logger, {
       'Accept': 'audio/mp3',
       'Content-Type': 'application/json'
     });
-    const mp3 = await post('/v1/rime-tts', {
+    const sampleRate = 8000;
+    const audioContent = await post('/v1/rime-tts', {
       speaker: voice,
       text,
       modelId: model_id,
-      samplingRate: 8000,
+      samplingRate: sampleRate,
       ...opts
     });
-    return mp3;
+    return {
+      audioContent,
+      extension: 'mp3',
+      sampleRate
+    };
   } catch (err) {
     logger.info({err}, 'synth rimelabs returned error');
     stats.increment('tts.count', ['vendor:rimelabs', 'accepted:no']);
@@ -919,13 +1000,17 @@ const synthVerbio = async(client, logger, {credentials, stats, voice, text, rend
       'User-Agent': 'jambonz',
       'Content-Type': 'application/json'
     });
-    const r8 = await post('/api/v1/synthesize', {
+    const audioContent = await post('/api/v1/synthesize', {
       voice_id: voice,
       output_sample_rate: '8k',
       output_encoding: 'pcm16',
       text
     });
-    return r8;
+    return {
+      audioContent,
+      extension: 'r8',
+      sampleRate: 8000
+    };
   } catch (err) {
     logger.info({err}, 'synth Verbio returned error');
     stats.increment('tts.count', ['vendor:verbio', 'accepted:no']);
@@ -965,7 +1050,11 @@ const synthWhisper = async(logger, {credentials, stats, voice, text, renderForCa
       input: text,
       response_format: 'mp3'
     });
-    return Buffer.from(await mp3.arrayBuffer());
+    return {
+      audioContent: Buffer.from(await mp3.arrayBuffer()),
+      extension: 'mp3',
+      sampleRate: 8000
+    };
   } catch (err) {
     logger.info({err}, 'synth whisper returned error');
     stats.increment('tts.count', ['vendor:openai', 'accepted:no']);
@@ -997,10 +1086,14 @@ const synthDeepgram = async(logger, {credentials, stats, model, text, renderForC
       'Accept': 'audio/mpeg',
       'Content-Type': 'application/json'
     });
-    const mp3 = await post(`/v1/speak?model=${model}`, {
+    const audioContent = await post(`/v1/speak?model=${model}`, {
       text
     });
-    return mp3;
+    return {
+      audioContent,
+      extension: 'mp3',
+      sampleRate: 8000
+    };
   } catch (err) {
     logger.info({err}, 'synth Deepgram returned error');
     stats.increment('tts.count', ['vendor:deepgram', 'accepted:no']);
@@ -1008,26 +1101,96 @@ const synthDeepgram = async(logger, {credentials, stats, model, text, renderForC
   }
 };
 
+const synthCartesia = async(logger, {
+  credentials, options, stats, voice, language, text, renderForCaching, disableTtsStreaming
+}) => {
+  const {api_key, model_id, embedding, options: credOpts} = credentials;
+  const opts = !!options && Object.keys(options).length !== 0 ? options : JSON.parse(credOpts || '{}');
+
+  if (!JAMBONES_DISABLE_TTS_STREAMING && !renderForCaching && !disableTtsStreaming) {
+    let params = '';
+    params += `{api_key=${api_key}`;
+    params += `,model_id=${model_id}`;
+    params += ',vendor=cartesia';
+    params += `,voice=${voice}`;
+    params += ',write_cache_file=1';
+    params += `,language=${language}`;
+    params += `,voice_mode=${embedding ? 'embedding' : 'id'}`;
+    if (embedding) params += `,embedding=${embedding}`;
+    if (opts.speed) params += `,speed=${opts.speed}`;
+    if (opts.emotion) params += `,emotion=${opts.emotion}`;
+    params += '}';
+
+    return {
+      filePath: `say:${params}${text.replace(/\n/g, ' ').replace(/\r/g, ' ')}`,
+      servedFromCache: false,
+      rtt: 0
+    };
+  }
+
+  try {
+    const client = new CartesiaClient({ apiKey: api_key });
+    const sampleRate = 48000;
+    const mp3 = await client.tts.bytes({
+      modelId: model_id,
+      transcript: text,
+      voice: {
+        mode: embedding ? 'embedding' : 'id',
+        ...(embedding ?
+          {
+            embedding: embedding.split(',').map(Number)
+          } :
+          {
+            id: voice
+          }
+        ),
+        ...(opts.speed || opts.emotion && {
+          experimentalControls: {
+            ...(opts.speed !== null && opts.speed !== undefined && {speed: opts.speed}),
+            ...(opts.emotion && {emotion: opts.emotion}),
+          }
+        })
+      },
+      language: language,
+      outputFormat: {
+        container: 'mp3',
+        bitRate: 128000,
+        sampleRate
+      },
+    });
+    return {
+      audioContent: Buffer.from(mp3),
+      extension: 'mp3',
+      sampleRate
+    };
+  } catch (err) {
+    logger.info({err}, 'synth Cartesia returned error');
+    stats.increment('tts.count', ['vendor:cartesia', 'accepted:no']);
+    throw err;
+  }
+
+};
+
 const getFileExtFromMime = (mime) => {
   switch (mime) {
     case 'audio/wav':
     case 'audio/x-wav':
-      return '.wav';
+      return ['wav', 8000];
     case /audio\/l16.*rate=8000/.test(mime) ? mime : 'cant match value':
-      return '.r8';
+      return ['r8', 8000];
     case /audio\/l16.*rate=16000/.test(mime) ? mime : 'cant match value':
-      return '.r16';
+      return ['r16', 16000];
     case /audio\/l16.*rate=24000/.test(mime) ? mime : 'cant match value':
-      return '.r24';
+      return ['r24', 24000];
     case /audio\/l16.*rate=32000/.test(mime) ? mime : 'cant match value':
-      return '.r32';
+      return ['r32', 32000];
     case /audio\/l16.*rate=48000/.test(mime) ? mime : 'cant match value':
-      return '.r48';
+      return ['r48', 48000];
     case 'audio/mpeg':
     case 'audio/mp3':
-      return '.mp3';
+      return ['mp3', 8000];
     default:
-      return '.wav';
+      return ['wav', 8000];
   }
 };
 

lib/utils.js

@@ -6,7 +6,7 @@ const pool = new Pool('https://auth.crt.nuance.com');
 const NUANCE_AUTH_ENDPOINT = 'tts.api.nuance.com:443';
 const grpc = require('@grpc/grpc-js');
 const formurlencoded = require('form-urlencoded');
-const { JAMBONES_DISABLE_TTS_STREAMING, JAMBONES_TTS_TRIM_SILENCE, TMP_FOLDER, HTTP_TIMEOUT } = require('./config');
+const { TMP_FOLDER, HTTP_TIMEOUT } = require('./config');
 
 const debug = require('debug')('jambonz:realtimedb-helpers');
 /**
@@ -17,59 +17,19 @@ const debug = require('debug')('jambonz:realtimedb-helpers');
 //const nuanceClientMap = new Map();
 
 function makeSynthKey({
-  account_sid = '', vendor, language, voice, engine = '', text,
-  renderForCaching = false}) {
+  account_sid = '', vendor, language, voice, engine = '', text}) {
   const hash = crypto.createHash('sha1');
   hash.update(`${language}:${vendor}:${voice}:${engine}:${text}`);
   const hexHashKey = hash.digest('hex');
   const accountKey = account_sid ? `:${account_sid}` : '';
-  const namespace = vendor.startsWith('custom') ? vendor : getFileExtension({vendor, renderForCaching});
-  const key = `tts${accountKey}:${namespace}:${hexHashKey}`;
+  const key = `tts${accountKey}:${hexHashKey}`;
   return key;
 }
 
-function makeFilePath({vendor, key, salt = '', renderForCaching = false}) {
-  const extension = getFileExtension({vendor, renderForCaching});
+function makeFilePath({key, salt = '', extension}) {
   return `${TMP_FOLDER}/${key.replace('tts:', `tts-${salt}`)}.${extension}`;
 }
 
-function getFileExtension({vendor, renderForCaching = false}) {
-  const mp3Extension = 'mp3';
-  const r8Extension = 'r8';
-
-  switch (vendor) {
-    case 'azure':
-    case 'microsoft':
-      if (!renderForCaching && !JAMBONES_DISABLE_TTS_STREAMING || JAMBONES_TTS_TRIM_SILENCE) {
-        return r8Extension;
-      } else {
-        return mp3Extension;
-      }
-    case 'deepgram':
-    case 'elevenlabs':
-    case 'rimelabs':
-    case 'playht':
-      if (renderForCaching || JAMBONES_DISABLE_TTS_STREAMING) {
-        return mp3Extension;
-      } else {
-        return r8Extension;
-      }
-    case 'nuance':
-    case 'nvidia':
-    case 'verbio':
-      return r8Extension;
-    default:
-      // If vendor is custom
-      if (vendor.startsWith('custom')) {
-        if (renderForCaching || JAMBONES_DISABLE_TTS_STREAMING) {
-          return mp3Extension;
-        } else {
-          return r8Extension;
-        }
-      }
-      return mp3Extension;
-  }
-}
 
 const noopLogger = {
   info: () => {},

package-lock.json

@@ -1,17 +1,18 @@
 {
   "name": "@jambonz/speech-utils",
-  "version": "0.1.20",
+  "version": "0.2.1",
   "lockfileVersion": 2,
   "requires": true,
   "packages": {
     "": {
       "name": "@jambonz/speech-utils",
-      "version": "0.1.20",
+      "version": "0.2.1",
       "license": "MIT",
       "dependencies": {
         "@aws-sdk/client-polly": "^3.496.0",
         "@aws-sdk/client-sts": "^3.496.0",
-        "@google-cloud/text-to-speech": "^5.0.2",
+        "@cartesia/cartesia-js": "^2.1.0",
+        "@google-cloud/text-to-speech": "^5.5.0",
         "@grpc/grpc-js": "^1.9.14",
         "@jambonz/realtimedb-helpers": "^0.8.7",
         "bent": "^7.3.12",
@@ -1173,6 +1174,92 @@
         "node": ">=6.9.0"
       }
     },
+    "node_modules/@cartesia/cartesia-js": {
+      "version": "2.1.0",
+      "resolved": "https://registry.npmjs.org/@cartesia/cartesia-js/-/cartesia-js-2.1.0.tgz",
+      "integrity": "sha512-r4uqhFWxVmFbm0A6aDeY0y3RVJHAc6PQ8hTwHmb2ZO2kmyAP4WoZdEEQ9GIU/5c1XpV5n3Y/y6zKjMzUwgG64A==",
+      "dependencies": {
+        "emittery": "^0.13.1",
+        "form-data": "^4.0.0",
+        "form-data-encoder": "^4.0.2",
+        "formdata-node": "^6.0.3",
+        "human-id": "^4.1.1",
+        "node-fetch": "2.7.0",
+        "qs": "6.11.2",
+        "readable-stream": "^4.5.2",
+        "url-join": "4.0.1",
+        "ws": "^8.15.13"
+      }
+    },
+    "node_modules/@cartesia/cartesia-js/node_modules/form-data": {
+      "version": "4.0.1",
+      "resolved": "https://registry.npmjs.org/form-data/-/form-data-4.0.1.tgz",
+      "integrity": "sha512-tzN8e4TX8+kkxGPK8D5u0FNmjPUjw3lwC9lSLxxoB/+GtsJG91CO8bSWy73APlgAZzZbXEYZJuxjkHH2w+Ezhw==",
+      "license": "MIT",
+      "dependencies": {
+        "asynckit": "^0.4.0",
+        "combined-stream": "^1.0.8",
+        "mime-types": "^2.1.12"
+      },
+      "engines": {
+        "node": ">= 6"
+      }
+    },
+    "node_modules/@cartesia/cartesia-js/node_modules/form-data-encoder": {
+      "version": "4.0.2",
+      "resolved": "https://registry.npmjs.org/form-data-encoder/-/form-data-encoder-4.0.2.tgz",
+      "integrity": "sha512-KQVhvhK8ZkWzxKxOr56CPulAhH3dobtuQ4+hNQ+HekH/Wp5gSOafqRAeTphQUJAIk0GBvHZgJ2ZGRWd5kphMuw==",
+      "license": "MIT",
+      "engines": {
+        "node": ">= 18"
+      }
+    },
+    "node_modules/@cartesia/cartesia-js/node_modules/formdata-node": {
+      "version": "6.0.3",
+      "resolved": "https://registry.npmjs.org/formdata-node/-/formdata-node-6.0.3.tgz",
+      "integrity": "sha512-8e1++BCiTzUno9v5IZ2J6bv4RU+3UKDmqWUQD0MIMVCd9AdhWkO1gw57oo1mNEX1dMq2EGI+FbWz4B92pscSQg==",
+      "license": "MIT",
+      "engines": {
+        "node": ">= 18"
+      }
+    },
+    "node_modules/@cartesia/cartesia-js/node_modules/readable-stream": {
+      "version": "4.5.2",
+      "resolved": "https://registry.npmjs.org/readable-stream/-/readable-stream-4.5.2.tgz",
+      "integrity": "sha512-yjavECdqeZ3GLXNgRXgeQEdz9fvDDkNKyHnbHRFtOr7/LcfgBcmct7t/ET+HaCTqfh06OzoAxrkN/IfjJBVe+g==",
+      "license": "MIT",
+      "dependencies": {
+        "abort-controller": "^3.0.0",
+        "buffer": "^6.0.3",
+        "events": "^3.3.0",
+        "process": "^0.11.10",
+        "string_decoder": "^1.3.0"
+      },
+      "engines": {
+        "node": "^12.22.0 || ^14.17.0 || >=16.0.0"
+      }
+    },
+    "node_modules/@cartesia/cartesia-js/node_modules/ws": {
+      "version": "8.18.0",
+      "resolved": "https://registry.npmjs.org/ws/-/ws-8.18.0.tgz",
+      "integrity": "sha512-8VbfWfHLbbwu3+N6OKsOMpBdT4kXPDDB9cJk2bJ6mh9ucxdlnNvH1e+roYkKmN9Nxw2yjz7VzeO9oOz2zJ04Pw==",
+      "license": "MIT",
+      "engines": {
+        "node": ">=10.0.0"
+      },
+      "peerDependencies": {
+        "bufferutil": "^4.0.1",
+        "utf-8-validate": ">=5.0.2"
+      },
+      "peerDependenciesMeta": {
+        "bufferutil": {
+          "optional": true
+        },
+        "utf-8-validate": {
+          "optional": true
+        }
+      }
+    },
     "node_modules/@eslint-community/eslint-utils": {
       "version": "4.4.0",
       "resolved": "https://registry.npmjs.org/@eslint-community/eslint-utils/-/eslint-utils-4.4.0.tgz",
@@ -1230,9 +1317,10 @@
       }
     },
     "node_modules/@google-cloud/text-to-speech": {
-      "version": "5.0.2",
-      "resolved": "https://registry.npmjs.org/@google-cloud/text-to-speech/-/text-to-speech-5.0.2.tgz",
-      "integrity": "sha512-Q11Ddh9eHKSDA3E/KSqMITgVprXb0XgIKuJP9F5ScJ1T9h+DNrbgIU7shd0QOlPqb8ruQRiTOqL08+Mq5R89Ow==",
+      "version": "5.5.0",
+      "resolved": "https://registry.npmjs.org/@google-cloud/text-to-speech/-/text-to-speech-5.5.0.tgz",
+      "integrity": "sha512-Cw/UK2Y3l31Vsuozu8cxsmVS/09fShimes0tRLgDbOY2ZMG1Dckb6Zf/Q3Nxg4X0feFep44pvwNmyHKrOnl9SQ==",
+      "license": "Apache-2.0",
       "dependencies": {
         "google-gax": "^4.0.3"
       },
@@ -2690,7 +2778,6 @@
       "version": "6.0.3",
       "resolved": "https://registry.npmjs.org/buffer/-/buffer-6.0.3.tgz",
       "integrity": "sha512-FTiCpNxtwiZZHEZbcbTIcZjERVICn9yq/pDFkTl95/AxzD1naBctN7YO68riM/gLSDY7sdrMby8hofADYuuqOA==",
-      "dev": true,
       "funding": [
         {
           "type": "github",
@@ -2751,7 +2838,6 @@
       "version": "1.0.7",
       "resolved": "https://registry.npmjs.org/call-bind/-/call-bind-1.0.7.tgz",
       "integrity": "sha512-GHTSNSYICQ7scH7sZ+M2rFopRoLh8t2bLSW6BbgrtLsahOIB5iyAVJf9GjWK3cYTDaMj4XdBpM1cA6pIS0Kv2w==",
-      "dev": true,
       "dependencies": {
         "es-define-property": "^1.0.0",
         "es-errors": "^1.3.0",
@@ -3034,7 +3120,6 @@
       "version": "1.1.4",
       "resolved": "https://registry.npmjs.org/define-data-property/-/define-data-property-1.1.4.tgz",
       "integrity": "sha512-rBMvIzlpA8v6E+SJZoo++HAYqsLrkg7MSfIinMPFhmkorw7X+dOXVJQs+QT69zGkzMyfDnIMN2Wid1+NbL3T+A==",
-      "dev": true,
       "dependencies": {
         "es-define-property": "^1.0.0",
         "es-errors": "^1.3.0",
@@ -3151,6 +3236,18 @@
       "integrity": "sha512-xT1HVAu5xFn7bDfkjGQi9dNpMqGchUkebwf1GL7cZN32NSwwlHRPMSDJ1KN6HkS0bWUtndbSQZqvpQftKG2uFQ==",
       "dev": true
     },
+    "node_modules/emittery": {
+      "version": "0.13.1",
+      "resolved": "https://registry.npmjs.org/emittery/-/emittery-0.13.1.tgz",
+      "integrity": "sha512-DeWwawk6r5yR9jFgnDKYt4sLS0LmHJJi3ZOnb5/JdbYwj3nW+FxQnHIjhBKz8YLC7oRNPVM9NQ47I3CVx34eqQ==",
+      "license": "MIT",
+      "engines": {
+        "node": ">=12"
+      },
+      "funding": {
+        "url": "https://github.com/sindresorhus/emittery?sponsor=1"
+      }
+    },
     "node_modules/emoji-regex": {
       "version": "8.0.0",
       "resolved": "https://registry.npmjs.org/emoji-regex/-/emoji-regex-8.0.0.tgz",
@@ -3221,7 +3318,6 @@
       "version": "1.0.0",
       "resolved": "https://registry.npmjs.org/es-define-property/-/es-define-property-1.0.0.tgz",
       "integrity": "sha512-jxayLKShrEqqzJ0eumQbVhTYQM27CfT1T35+gCgDFoL82JLsXqTJ76zv6A0YLOgEnLUMvLzsDsGIrl8NFpT2gQ==",
-      "dev": true,
       "dependencies": {
         "get-intrinsic": "^1.2.4"
       },
@@ -3233,7 +3329,6 @@
       "version": "1.3.0",
       "resolved": "https://registry.npmjs.org/es-errors/-/es-errors-1.3.0.tgz",
       "integrity": "sha512-Zf5H2Kxt2xjTvbJvP2ZWLEICxA6j+hAmMzIlypy4xcBg1vKVnx89Wy0GbS+kf5cwCVFFzdCFh2XSCFNULS6csw==",
-      "dev": true,
       "engines": {
         "node": ">= 0.4"
       }
@@ -3579,7 +3674,6 @@
       "version": "3.3.0",
       "resolved": "https://registry.npmjs.org/events/-/events-3.3.0.tgz",
       "integrity": "sha512-mQw+2fkQbALzQ7V0MY0IqdnXNOeTtP4r0lN9z7AAawCXgqea7bDii20AYrIBrFd/Hx0M2Ocz6S111CaFkUcb0Q==",
-      "dev": true,
       "engines": {
         "node": ">=0.8.x"
       }
@@ -3881,7 +3975,6 @@
       "version": "1.1.2",
       "resolved": "https://registry.npmjs.org/function-bind/-/function-bind-1.1.2.tgz",
       "integrity": "sha512-7XHNxH7qX9xG5mIwxkhumTox/MIRNcOgDrxWsMt2pAr23WHp6MrRlN7FBSFpCpr+oVO0F744iUgR82nJMfG2SA==",
-      "dev": true,
       "funding": {
         "url": "https://github.com/sponsors/ljharb"
       }
@@ -3960,7 +4053,6 @@
       "version": "1.2.4",
       "resolved": "https://registry.npmjs.org/get-intrinsic/-/get-intrinsic-1.2.4.tgz",
       "integrity": "sha512-5uYhsJH8VJBTv7oslg4BznJYhDoRI6waYCxMmCdnTrcCrHA/fCFKoTFz2JKKE0HdDFUF7/oQuhzumXJK7paBRQ==",
-      "dev": true,
       "dependencies": {
         "es-errors": "^1.3.0",
         "function-bind": "^1.1.2",
@@ -4105,7 +4197,6 @@
       "version": "1.0.1",
       "resolved": "https://registry.npmjs.org/gopd/-/gopd-1.0.1.tgz",
       "integrity": "sha512-d65bNlIadxvpb/A2abVdlqKqV563juRnZ1Wtk6s1sIR8uNsXR70xqIzVqxVf1eTqDunwT2MkczEeaezCKTZhwA==",
-      "dev": true,
       "dependencies": {
         "get-intrinsic": "^1.1.3"
       },
@@ -4167,7 +4258,6 @@
       "version": "1.0.2",
       "resolved": "https://registry.npmjs.org/has-property-descriptors/-/has-property-descriptors-1.0.2.tgz",
       "integrity": "sha512-55JNKuIW+vq4Ke1BjOTjM2YctQIvCT7GFzHwmfZPGo5wnrgkid0YQtnAleFSqumZm4az3n2BS+erby5ipJdgrg==",
-      "dev": true,
       "dependencies": {
         "es-define-property": "^1.0.0"
       },
@@ -4179,7 +4269,6 @@
       "version": "1.0.1",
       "resolved": "https://registry.npmjs.org/has-proto/-/has-proto-1.0.1.tgz",
       "integrity": "sha512-7qE+iP+O+bgF9clE5+UoBFzE65mlBiVj3tKCrlNQ0Ogwm0BjpT/gK4SlLYDMybDh5I3TCTKnPPa0oMG7JDYrhg==",
-      "dev": true,
       "engines": {
         "node": ">= 0.4"
       },
@@ -4191,7 +4280,6 @@
       "version": "1.0.3",
       "resolved": "https://registry.npmjs.org/has-symbols/-/has-symbols-1.0.3.tgz",
       "integrity": "sha512-l3LCuF6MgDNwTDKkdYGEihYjt5pRPbEg46rtlmnSPlUbgmB8LOIrKJbYYFBSbnPaJexMKtiPO8hmeRjRz2Td+A==",
-      "dev": true,
       "engines": {
         "node": ">= 0.4"
       },
@@ -4243,7 +4331,6 @@
       "version": "2.0.2",
       "resolved": "https://registry.npmjs.org/hasown/-/hasown-2.0.2.tgz",
       "integrity": "sha512-0hJU9SCPvmMzIBdZFqNPXWa6dqh7WdH0cII9y+CyS8rG3nL48Bclra9HmKhVVUHyPWNH5Y7xDwAB7bfgSjkUMQ==",
-      "dev": true,
       "dependencies": {
         "function-bind": "^1.1.2"
       },
@@ -4293,6 +4380,15 @@
         "node": ">= 14"
       }
     },
+    "node_modules/human-id": {
+      "version": "4.1.1",
+      "resolved": "https://registry.npmjs.org/human-id/-/human-id-4.1.1.tgz",
+      "integrity": "sha512-3gKm/gCSUipeLsRYZbbdA1BD83lBoWUkZ7G9VFrhWPAU76KwYo5KR8V28bpoPm/ygy0x5/GCbpRQdY7VLYCoIg==",
+      "license": "MIT",
+      "bin": {
+        "human-id": "dist/cli.js"
+      }
+    },
     "node_modules/humanize-ms": {
       "version": "1.2.1",
       "resolved": "https://registry.npmjs.org/humanize-ms/-/humanize-ms-1.2.1.tgz",
@@ -5689,7 +5785,6 @@
       "version": "1.13.1",
       "resolved": "https://registry.npmjs.org/object-inspect/-/object-inspect-1.13.1.tgz",
       "integrity": "sha512-5qoj1RUiKOMsCCNLV1CBiPYE10sziTsnmNxkAI/rZhiD63CF7IqdFGC/XzjWjpSgLf0LxXX3bDFIh0E18f6UhQ==",
-      "dev": true,
       "funding": {
         "url": "https://github.com/sponsors/ljharb"
       }
@@ -6083,7 +6178,6 @@
       "version": "0.11.10",
       "resolved": "https://registry.npmjs.org/process/-/process-0.11.10.tgz",
       "integrity": "sha512-cdGef/drWFoydD1JsMzuFf8100nZl+GT+yacc2bEced5f9Rjk4z+WtFUTBu9PhOi9j/jfmBPu0mMEY4wIdAF8A==",
-      "dev": true,
       "engines": {
         "node": ">= 0.6.0"
       }
@@ -6158,6 +6252,21 @@
         "node": ">=6"
       }
     },
+    "node_modules/qs": {
+      "version": "6.11.2",
+      "resolved": "https://registry.npmjs.org/qs/-/qs-6.11.2.tgz",
+      "integrity": "sha512-tDNIz22aBzCDxLtVH++VnTfzxlfeK5CbqohpSqpJgj1Wg/cQbStNAz3NuqCs5vV+pjBsK4x4pN9HlVh7rcYRiA==",
+      "license": "BSD-3-Clause",
+      "dependencies": {
+        "side-channel": "^1.0.4"
+      },
+      "engines": {
+        "node": ">=0.6"
+      },
+      "funding": {
+        "url": "https://github.com/sponsors/ljharb"
+      }
+    },
     "node_modules/querystringify": {
       "version": "2.2.0",
       "resolved": "https://registry.npmjs.org/querystringify/-/querystringify-2.2.0.tgz",
@@ -6480,7 +6589,6 @@
       "version": "1.2.2",
       "resolved": "https://registry.npmjs.org/set-function-length/-/set-function-length-1.2.2.tgz",
       "integrity": "sha512-pgRc4hJ4/sNjWCSS9AmnS40x3bNMDTknHgL5UaMBTMyJnU90EgWh1Rz+MC9eFu4BuN/UwZjKQuY/1v3rM7HMfg==",
-      "dev": true,
       "dependencies": {
         "define-data-property": "^1.1.4",
         "es-errors": "^1.3.0",
@@ -6532,7 +6640,6 @@
       "version": "1.0.4",
       "resolved": "https://registry.npmjs.org/side-channel/-/side-channel-1.0.4.tgz",
       "integrity": "sha512-q5XPytqFEIKHkGdiMIrY10mvLRvnQh42/+GoBlFW3b2LXLE2xxJpZFdm94we0BaoV3RwJyGqg5wS7epxTv0Zvw==",
-      "dev": true,
       "dependencies": {
         "call-bind": "^1.0.0",
         "get-intrinsic": "^1.0.2",
@@ -7132,6 +7239,12 @@
         "punycode": "^2.1.0"
       }
     },
+    "node_modules/url-join": {
+      "version": "4.0.1",
+      "resolved": "https://registry.npmjs.org/url-join/-/url-join-4.0.1.tgz",
+      "integrity": "sha512-jk1+QP6ZJqyOiuEI9AEWQfju/nB2Pw466kbA0LEZljHwKeMgd9WrAEgEGxjPDD2+TNbbb37rTyhEfrCXfuKXnA==",
+      "license": "MIT"
+    },
     "node_modules/url-parse": {
       "version": "1.5.10",
       "resolved": "https://registry.npmjs.org/url-parse/-/url-parse-1.5.10.tgz",
@@ -8299,6 +8412,63 @@
         "to-fast-properties": "^2.0.0"
       }
     },
+    "@cartesia/cartesia-js": {
+      "version": "2.1.0",
+      "resolved": "https://registry.npmjs.org/@cartesia/cartesia-js/-/cartesia-js-2.1.0.tgz",
+      "integrity": "sha512-r4uqhFWxVmFbm0A6aDeY0y3RVJHAc6PQ8hTwHmb2ZO2kmyAP4WoZdEEQ9GIU/5c1XpV5n3Y/y6zKjMzUwgG64A==",
+      "requires": {
+        "emittery": "^0.13.1",
+        "form-data": "^4.0.0",
+        "form-data-encoder": "^4.0.2",
+        "formdata-node": "^6.0.3",
+        "human-id": "^4.1.1",
+        "node-fetch": "2.7.0",
+        "qs": "6.11.2",
+        "readable-stream": "^4.5.2",
+        "url-join": "4.0.1",
+        "ws": "^8.15.13"
+      },
+      "dependencies": {
+        "form-data": {
+          "version": "4.0.1",
+          "resolved": "https://registry.npmjs.org/form-data/-/form-data-4.0.1.tgz",
+          "integrity": "sha512-tzN8e4TX8+kkxGPK8D5u0FNmjPUjw3lwC9lSLxxoB/+GtsJG91CO8bSWy73APlgAZzZbXEYZJuxjkHH2w+Ezhw==",
+          "requires": {
+            "asynckit": "^0.4.0",
+            "combined-stream": "^1.0.8",
+            "mime-types": "^2.1.12"
+          }
+        },
+        "form-data-encoder": {
+          "version": "4.0.2",
+          "resolved": "https://registry.npmjs.org/form-data-encoder/-/form-data-encoder-4.0.2.tgz",
+          "integrity": "sha512-KQVhvhK8ZkWzxKxOr56CPulAhH3dobtuQ4+hNQ+HekH/Wp5gSOafqRAeTphQUJAIk0GBvHZgJ2ZGRWd5kphMuw=="
+        },
+        "formdata-node": {
+          "version": "6.0.3",
+          "resolved": "https://registry.npmjs.org/formdata-node/-/formdata-node-6.0.3.tgz",
+          "integrity": "sha512-8e1++BCiTzUno9v5IZ2J6bv4RU+3UKDmqWUQD0MIMVCd9AdhWkO1gw57oo1mNEX1dMq2EGI+FbWz4B92pscSQg=="
+        },
+        "readable-stream": {
+          "version": "4.5.2",
+          "resolved": "https://registry.npmjs.org/readable-stream/-/readable-stream-4.5.2.tgz",
+          "integrity": "sha512-yjavECdqeZ3GLXNgRXgeQEdz9fvDDkNKyHnbHRFtOr7/LcfgBcmct7t/ET+HaCTqfh06OzoAxrkN/IfjJBVe+g==",
+          "requires": {
+            "abort-controller": "^3.0.0",
+            "buffer": "^6.0.3",
+            "events": "^3.3.0",
+            "process": "^0.11.10",
+            "string_decoder": "^1.3.0"
+          }
+        },
+        "ws": {
+          "version": "8.18.0",
+          "resolved": "https://registry.npmjs.org/ws/-/ws-8.18.0.tgz",
+          "integrity": "sha512-8VbfWfHLbbwu3+N6OKsOMpBdT4kXPDDB9cJk2bJ6mh9ucxdlnNvH1e+roYkKmN9Nxw2yjz7VzeO9oOz2zJ04Pw==",
+          "requires": {}
+        }
+      }
+    },
     "@eslint-community/eslint-utils": {
       "version": "4.4.0",
       "resolved": "https://registry.npmjs.org/@eslint-community/eslint-utils/-/eslint-utils-4.4.0.tgz",
@@ -8338,9 +8508,9 @@
       "dev": true
     },
     "@google-cloud/text-to-speech": {
-      "version": "5.0.2",
-      "resolved": "https://registry.npmjs.org/@google-cloud/text-to-speech/-/text-to-speech-5.0.2.tgz",
-      "integrity": "sha512-Q11Ddh9eHKSDA3E/KSqMITgVprXb0XgIKuJP9F5ScJ1T9h+DNrbgIU7shd0QOlPqb8ruQRiTOqL08+Mq5R89Ow==",
+      "version": "5.5.0",
+      "resolved": "https://registry.npmjs.org/@google-cloud/text-to-speech/-/text-to-speech-5.5.0.tgz",
+      "integrity": "sha512-Cw/UK2Y3l31Vsuozu8cxsmVS/09fShimes0tRLgDbOY2ZMG1Dckb6Zf/Q3Nxg4X0feFep44pvwNmyHKrOnl9SQ==",
       "requires": {
         "google-gax": "^4.0.3"
       }
@@ -9458,7 +9628,6 @@
       "version": "6.0.3",
       "resolved": "https://registry.npmjs.org/buffer/-/buffer-6.0.3.tgz",
       "integrity": "sha512-FTiCpNxtwiZZHEZbcbTIcZjERVICn9yq/pDFkTl95/AxzD1naBctN7YO68riM/gLSDY7sdrMby8hofADYuuqOA==",
-      "dev": true,
       "requires": {
         "base64-js": "^1.3.1",
         "ieee754": "^1.2.1"
@@ -9498,7 +9667,6 @@
       "version": "1.0.7",
       "resolved": "https://registry.npmjs.org/call-bind/-/call-bind-1.0.7.tgz",
       "integrity": "sha512-GHTSNSYICQ7scH7sZ+M2rFopRoLh8t2bLSW6BbgrtLsahOIB5iyAVJf9GjWK3cYTDaMj4XdBpM1cA6pIS0Kv2w==",
-      "dev": true,
       "requires": {
         "es-define-property": "^1.0.0",
         "es-errors": "^1.3.0",
@@ -9696,7 +9864,6 @@
       "version": "1.1.4",
       "resolved": "https://registry.npmjs.org/define-data-property/-/define-data-property-1.1.4.tgz",
       "integrity": "sha512-rBMvIzlpA8v6E+SJZoo++HAYqsLrkg7MSfIinMPFhmkorw7X+dOXVJQs+QT69zGkzMyfDnIMN2Wid1+NbL3T+A==",
-      "dev": true,
       "requires": {
         "es-define-property": "^1.0.0",
         "es-errors": "^1.3.0",
@@ -9783,6 +9950,11 @@
       "integrity": "sha512-xT1HVAu5xFn7bDfkjGQi9dNpMqGchUkebwf1GL7cZN32NSwwlHRPMSDJ1KN6HkS0bWUtndbSQZqvpQftKG2uFQ==",
       "dev": true
     },
+    "emittery": {
+      "version": "0.13.1",
+      "resolved": "https://registry.npmjs.org/emittery/-/emittery-0.13.1.tgz",
+      "integrity": "sha512-DeWwawk6r5yR9jFgnDKYt4sLS0LmHJJi3ZOnb5/JdbYwj3nW+FxQnHIjhBKz8YLC7oRNPVM9NQ47I3CVx34eqQ=="
+    },
     "emoji-regex": {
       "version": "8.0.0",
       "resolved": "https://registry.npmjs.org/emoji-regex/-/emoji-regex-8.0.0.tgz",
@@ -9847,7 +10019,6 @@
       "version": "1.0.0",
       "resolved": "https://registry.npmjs.org/es-define-property/-/es-define-property-1.0.0.tgz",
       "integrity": "sha512-jxayLKShrEqqzJ0eumQbVhTYQM27CfT1T35+gCgDFoL82JLsXqTJ76zv6A0YLOgEnLUMvLzsDsGIrl8NFpT2gQ==",
-      "dev": true,
       "requires": {
         "get-intrinsic": "^1.2.4"
       }
@@ -9855,8 +10026,7 @@
     "es-errors": {
       "version": "1.3.0",
       "resolved": "https://registry.npmjs.org/es-errors/-/es-errors-1.3.0.tgz",
-      "integrity": "sha512-Zf5H2Kxt2xjTvbJvP2ZWLEICxA6j+hAmMzIlypy4xcBg1vKVnx89Wy0GbS+kf5cwCVFFzdCFh2XSCFNULS6csw==",
-      "dev": true
+      "integrity": "sha512-Zf5H2Kxt2xjTvbJvP2ZWLEICxA6j+hAmMzIlypy4xcBg1vKVnx89Wy0GbS+kf5cwCVFFzdCFh2XSCFNULS6csw=="
     },
     "es-get-iterator": {
       "version": "1.1.3",
@@ -10107,8 +10277,7 @@
     "events": {
       "version": "3.3.0",
       "resolved": "https://registry.npmjs.org/events/-/events-3.3.0.tgz",
-      "integrity": "sha512-mQw+2fkQbALzQ7V0MY0IqdnXNOeTtP4r0lN9z7AAawCXgqea7bDii20AYrIBrFd/Hx0M2Ocz6S111CaFkUcb0Q==",
-      "dev": true
+      "integrity": "sha512-mQw+2fkQbALzQ7V0MY0IqdnXNOeTtP4r0lN9z7AAawCXgqea7bDii20AYrIBrFd/Hx0M2Ocz6S111CaFkUcb0Q=="
     },
     "expect": {
       "version": "26.6.2",
@@ -10323,8 +10492,7 @@
     "function-bind": {
       "version": "1.1.2",
       "resolved": "https://registry.npmjs.org/function-bind/-/function-bind-1.1.2.tgz",
-      "integrity": "sha512-7XHNxH7qX9xG5mIwxkhumTox/MIRNcOgDrxWsMt2pAr23WHp6MrRlN7FBSFpCpr+oVO0F744iUgR82nJMfG2SA==",
-      "dev": true
+      "integrity": "sha512-7XHNxH7qX9xG5mIwxkhumTox/MIRNcOgDrxWsMt2pAr23WHp6MrRlN7FBSFpCpr+oVO0F744iUgR82nJMfG2SA=="
     },
     "function.prototype.name": {
       "version": "1.1.6",
@@ -10379,7 +10547,6 @@
       "version": "1.2.4",
       "resolved": "https://registry.npmjs.org/get-intrinsic/-/get-intrinsic-1.2.4.tgz",
       "integrity": "sha512-5uYhsJH8VJBTv7oslg4BznJYhDoRI6waYCxMmCdnTrcCrHA/fCFKoTFz2JKKE0HdDFUF7/oQuhzumXJK7paBRQ==",
-      "dev": true,
       "requires": {
         "es-errors": "^1.3.0",
         "function-bind": "^1.1.2",
@@ -10482,7 +10649,6 @@
       "version": "1.0.1",
       "resolved": "https://registry.npmjs.org/gopd/-/gopd-1.0.1.tgz",
       "integrity": "sha512-d65bNlIadxvpb/A2abVdlqKqV563juRnZ1Wtk6s1sIR8uNsXR70xqIzVqxVf1eTqDunwT2MkczEeaezCKTZhwA==",
-      "dev": true,
       "requires": {
         "get-intrinsic": "^1.1.3"
       }
@@ -10526,7 +10692,6 @@
       "version": "1.0.2",
       "resolved": "https://registry.npmjs.org/has-property-descriptors/-/has-property-descriptors-1.0.2.tgz",
       "integrity": "sha512-55JNKuIW+vq4Ke1BjOTjM2YctQIvCT7GFzHwmfZPGo5wnrgkid0YQtnAleFSqumZm4az3n2BS+erby5ipJdgrg==",
-      "dev": true,
       "requires": {
         "es-define-property": "^1.0.0"
       }
@@ -10534,14 +10699,12 @@
     "has-proto": {
       "version": "1.0.1",
       "resolved": "https://registry.npmjs.org/has-proto/-/has-proto-1.0.1.tgz",
-      "integrity": "sha512-7qE+iP+O+bgF9clE5+UoBFzE65mlBiVj3tKCrlNQ0Ogwm0BjpT/gK4SlLYDMybDh5I3TCTKnPPa0oMG7JDYrhg==",
-      "dev": true
+      "integrity": "sha512-7qE+iP+O+bgF9clE5+UoBFzE65mlBiVj3tKCrlNQ0Ogwm0BjpT/gK4SlLYDMybDh5I3TCTKnPPa0oMG7JDYrhg=="
     },
     "has-symbols": {
       "version": "1.0.3",
       "resolved": "https://registry.npmjs.org/has-symbols/-/has-symbols-1.0.3.tgz",
-      "integrity": "sha512-l3LCuF6MgDNwTDKkdYGEihYjt5pRPbEg46rtlmnSPlUbgmB8LOIrKJbYYFBSbnPaJexMKtiPO8hmeRjRz2Td+A==",
-      "dev": true
+      "integrity": "sha512-l3LCuF6MgDNwTDKkdYGEihYjt5pRPbEg46rtlmnSPlUbgmB8LOIrKJbYYFBSbnPaJexMKtiPO8hmeRjRz2Td+A=="
     },
     "has-tostringtag": {
       "version": "1.0.0",
@@ -10574,7 +10737,6 @@
       "version": "2.0.2",
       "resolved": "https://registry.npmjs.org/hasown/-/hasown-2.0.2.tgz",
       "integrity": "sha512-0hJU9SCPvmMzIBdZFqNPXWa6dqh7WdH0cII9y+CyS8rG3nL48Bclra9HmKhVVUHyPWNH5Y7xDwAB7bfgSjkUMQ==",
-      "dev": true,
       "requires": {
         "function-bind": "^1.1.2"
       }
@@ -10614,6 +10776,11 @@
         }
       }
     },
+    "human-id": {
+      "version": "4.1.1",
+      "resolved": "https://registry.npmjs.org/human-id/-/human-id-4.1.1.tgz",
+      "integrity": "sha512-3gKm/gCSUipeLsRYZbbdA1BD83lBoWUkZ7G9VFrhWPAU76KwYo5KR8V28bpoPm/ygy0x5/GCbpRQdY7VLYCoIg=="
+    },
     "humanize-ms": {
       "version": "1.2.1",
       "resolved": "https://registry.npmjs.org/humanize-ms/-/humanize-ms-1.2.1.tgz",
@@ -11668,8 +11835,7 @@
     "object-inspect": {
       "version": "1.13.1",
       "resolved": "https://registry.npmjs.org/object-inspect/-/object-inspect-1.13.1.tgz",
-      "integrity": "sha512-5qoj1RUiKOMsCCNLV1CBiPYE10sziTsnmNxkAI/rZhiD63CF7IqdFGC/XzjWjpSgLf0LxXX3bDFIh0E18f6UhQ==",
-      "dev": true
+      "integrity": "sha512-5qoj1RUiKOMsCCNLV1CBiPYE10sziTsnmNxkAI/rZhiD63CF7IqdFGC/XzjWjpSgLf0LxXX3bDFIh0E18f6UhQ=="
     },
     "object-is": {
       "version": "1.1.5",
@@ -11965,8 +12131,7 @@
     "process": {
       "version": "0.11.10",
       "resolved": "https://registry.npmjs.org/process/-/process-0.11.10.tgz",
-      "integrity": "sha512-cdGef/drWFoydD1JsMzuFf8100nZl+GT+yacc2bEced5f9Rjk4z+WtFUTBu9PhOi9j/jfmBPu0mMEY4wIdAF8A==",
-      "dev": true
+      "integrity": "sha512-cdGef/drWFoydD1JsMzuFf8100nZl+GT+yacc2bEced5f9Rjk4z+WtFUTBu9PhOi9j/jfmBPu0mMEY4wIdAF8A=="
     },
     "process-on-spawn": {
       "version": "1.0.0",
@@ -12025,6 +12190,14 @@
       "resolved": "https://registry.npmjs.org/punycode/-/punycode-2.3.1.tgz",
       "integrity": "sha512-vYt7UD1U9Wg6138shLtLOvdAu+8DsC/ilFtEVHcH+wydcSpNE20AfSOduf6MkRFahL5FY7X1oU7nKVZFtfq8Fg=="
     },
+    "qs": {
+      "version": "6.11.2",
+      "resolved": "https://registry.npmjs.org/qs/-/qs-6.11.2.tgz",
+      "integrity": "sha512-tDNIz22aBzCDxLtVH++VnTfzxlfeK5CbqohpSqpJgj1Wg/cQbStNAz3NuqCs5vV+pjBsK4x4pN9HlVh7rcYRiA==",
+      "requires": {
+        "side-channel": "^1.0.4"
+      }
+    },
     "querystringify": {
       "version": "2.2.0",
       "resolved": "https://registry.npmjs.org/querystringify/-/querystringify-2.2.0.tgz",
@@ -12229,7 +12402,6 @@
       "version": "1.2.2",
       "resolved": "https://registry.npmjs.org/set-function-length/-/set-function-length-1.2.2.tgz",
       "integrity": "sha512-pgRc4hJ4/sNjWCSS9AmnS40x3bNMDTknHgL5UaMBTMyJnU90EgWh1Rz+MC9eFu4BuN/UwZjKQuY/1v3rM7HMfg==",
-      "dev": true,
       "requires": {
         "define-data-property": "^1.1.4",
         "es-errors": "^1.3.0",
@@ -12269,7 +12441,6 @@
       "version": "1.0.4",
       "resolved": "https://registry.npmjs.org/side-channel/-/side-channel-1.0.4.tgz",
       "integrity": "sha512-q5XPytqFEIKHkGdiMIrY10mvLRvnQh42/+GoBlFW3b2LXLE2xxJpZFdm94we0BaoV3RwJyGqg5wS7epxTv0Zvw==",
-      "dev": true,
       "requires": {
         "call-bind": "^1.0.0",
         "get-intrinsic": "^1.0.2",
@@ -12717,6 +12888,11 @@
         "punycode": "^2.1.0"
       }
     },
+    "url-join": {
+      "version": "4.0.1",
+      "resolved": "https://registry.npmjs.org/url-join/-/url-join-4.0.1.tgz",
+      "integrity": "sha512-jk1+QP6ZJqyOiuEI9AEWQfju/nB2Pw466kbA0LEZljHwKeMgd9WrAEgEGxjPDD2+TNbbb37rTyhEfrCXfuKXnA=="
+    },
     "url-parse": {
       "version": "1.5.10",
       "resolved": "https://registry.npmjs.org/url-parse/-/url-parse-1.5.10.tgz",

package.json

@@ -1,6 +1,6 @@
 {
   "name": "@jambonz/speech-utils",
-  "version": "0.1.20",
+  "version": "0.2.1",
   "description": "TTS-related speech utilities for jambonz",
   "main": "index.js",
   "author": "Dave Horton",
@@ -28,7 +28,8 @@
   "dependencies": {
     "@aws-sdk/client-polly": "^3.496.0",
     "@aws-sdk/client-sts": "^3.496.0",
-    "@google-cloud/text-to-speech": "^5.0.2",
+    "@cartesia/cartesia-js": "^2.1.0",
+    "@google-cloud/text-to-speech": "^5.5.0",
     "@grpc/grpc-js": "^1.9.14",
     "@jambonz/realtimedb-helpers": "^0.8.7",
     "bent": "^7.3.12",

test/synth.js

@@ -91,14 +91,17 @@ test('Google speech Custom voice synth tests', async(t) => {
   const fn = require('..');
   const {synthAudio, client} = fn(opts, logger);
 
-  if (!process.env.GCP_CUSTOM_VOICE_FILE && !process.env.GCP_CUSTOM_VOICE_JSON_KEY || !process.env.GCP_CUSTOM_VOICE_MODEL) {
-    t.pass('skipping google speech synth tests since neither GCP_CUSTOM_VOICE_FILE nor GCP_CUSTOM_VOICE_JSON_KEY provided, GCP_CUSTOM_VOICE_MODEL is not provided');
+  if (!process.env.GCP_CUSTOM_VOICE_FILE &&
+    !process.env.GCP_CUSTOM_VOICE_JSON_KEY ||
+    !process.env.GCP_CUSTOM_VOICE_MODEL) {
+    t.pass(`skipping google speech synth tests since neither 
+GCP_CUSTOM_VOICE_FILE nor GCP_CUSTOM_VOICE_JSON_KEY provided, GCP_CUSTOM_VOICE_MODEL is not provided`);
     return t.end();
   }
   try {
     const str = process.env.GCP_CUSTOM_VOICE_JSON_KEY || fs.readFileSync(process.env.GCP_CUSTOM_VOICE_FILE);
     const creds = JSON.parse(str);
-    let opts = await synthAudio(stats, {
+    const opts = await synthAudio(stats, {
       vendor: 'google',
       credentials: {
         credentials: {
@@ -109,7 +112,7 @@ test('Google speech Custom voice synth tests', async(t) => {
       language: 'en-AU',
       text: 'This is a test.  This is only a test',
       voice: {
-        reportedUsage:"REALTIME",
+        reportedUsage: 'REALTIME',
         model: process.env.GCP_CUSTOM_VOICE_MODEL
       }
     });
@@ -121,6 +124,48 @@ test('Google speech Custom voice synth tests', async(t) => {
   client.quit();
 });
 
+test('Google speech voice cloning synth tests', async(t) => {
+  const fn = require('..');
+  const {synthAudio, client} = fn(opts, logger);
+
+  if (!process.env.GCP_CUSTOM_VOICE_FILE &&
+    !process.env.GCP_CUSTOM_VOICE_JSON_KEY ||
+    !process.env.GCP_VOICE_CLONING_FILE &&
+    !process.env.GCP_VOICE_CLONING_JSON_KEY) {
+    t.pass(`skipping google speech synth tests since neither 
+GCP_CUSTOM_VOICE_FILE nor GCP_CUSTOM_VOICE_JSON_KEY provided,
+GCP_VOICE_CLONING_FILE nor GCP_VOICE_CLONING_JSON_KEY is not provided`);
+    return t.end();
+  }
+  try {
+    const googleKey = process.env.GCP_CUSTOM_VOICE_JSON_KEY ||
+      fs.readFileSync(process.env.GCP_CUSTOM_VOICE_FILE);
+    const voice_cloning_key = process.env.GCP_VOICE_CLONING_JSON_KEY ||
+      fs.readFileSync(process.env.GCP_VOICE_CLONING_FILE).toString();
+    const creds = JSON.parse(googleKey);
+    const opts = await synthAudio(stats, {
+      vendor: 'google',
+      credentials: {
+        credentials: {
+          client_email: creds.client_email,
+          private_key: creds.private_key,
+          project_id: creds.project_id
+        },
+      },
+      language: 'en-US',
+      text: 'This is a test. This is only a test. This is a test. This is only a test. This is a test. This is only a test',
+      voice: {
+        voice_cloning_key
+      }
+    });
+    t.ok(!opts.servedFromCache, `successfully synthesized google voice cloning audio to ${opts.filePath}`);
+  } catch (err) {
+    console.error(err);
+    t.end(err);
+  }
+  client.quit();
+});
+
 test('AWS speech synth tests', async(t) => {
   const fn = require('..');
   const {synthAudio, client} = fn(opts, logger);
@@ -509,6 +554,7 @@ test('Custom Vendor speech synth tests', async(t) => {
       text: 'This is a test.  This is only a test',
     });
     t.ok(!opts.servedFromCache, `successfully synthesized custom vendor audio to ${opts.filePath}`);
+    t.ok(opts.filePath.endsWith('wav'), 'audio is cached as wav file');
     let obj = await getJSON(`http://127.0.0.1:3100/lastRequest/somethingnew`);
     t.ok(obj.headers.Authorization == 'Bearer some_jwt_token', 'Custom Vendor Authentication Header is correct');
     t.ok(obj.body.language == 'en-US', 'Custom Vendor Language is correct');
@@ -516,6 +562,21 @@ test('Custom Vendor speech synth tests', async(t) => {
     t.ok(obj.body.type == 'text', 'Custom Vendor type is correct');
     t.ok(obj.body.text == 'This is a test.  This is only a test', 'Custom Vendor text is correct');
 
+    // Checking if cache is stored with wav format
+    opts = await synthAudio(stats, {
+      vendor: 'custom:somethingnew',
+      credentials: {
+        use_for_tts: 1,
+        custom_tts_url: "http://127.0.0.1:3100/somethingnew",
+        auth_token: 'some_jwt_token'
+      },
+      language: 'en-US',
+      voice: 'English-US.Female-1',
+      text: 'This is a test.  This is only a test',
+    });
+    t.ok(opts.servedFromCache, `successfully get custom vendor cached audio to ${opts.filePath}`);
+    t.ok(opts.filePath.endsWith('wav'), 'audio is cached as wav file');
+
     opts = await synthAudio(stats, {
       vendor: 'custom:somethingnew2',
       credentials: {
@@ -625,6 +686,40 @@ test('PlayHT3.0 speech synth tests', async(t) => {
   await testPlayHT(t, 'Play3.0');
 });
 
+test('Cartesia speech synth tests', async(t) => {
+  const fn = require('..');
+  const {synthAudio, client} = fn(opts, logger);
+
+  if (!process.env.CARTESIA_API_KEY) {
+    t.pass('skipping Cartesia speech synth tests since CARTESIA_API_KEY is not provided');
+    return t.end();
+  }
+  const text = 'Hi there and welcome to jambones! ' + Date.now();
+  try {
+    const opts = await synthAudio(stats, {
+      vendor: 'cartesia',
+      credentials: {
+        api_key: process.env.CARTESIA_API_KEY,
+        model_id: 'sonic-english',
+        options: JSON.stringify({
+          speed: 1,
+          emotion: 'female_happy',
+        })
+      },
+      language: 'en',
+      voice: '694f9389-aac1-45b6-b726-9d9369183238',
+      text,
+      renderForCaching: true
+    });
+    t.ok(!opts.servedFromCache, `successfully playht eleven audio to ${opts.filePath}`);
+
+  } catch (err) {
+    console.error(JSON.stringify(err));
+    t.end(err);
+  }
+  client.quit();
+});
+
 test('rimelabs speech synth tests', async(t) => {
   const fn = require('..');
   const {synthAudio, client} = fn(opts, logger);
@@ -793,24 +888,15 @@ test('TTS Cache tests', async(t) => {
       language: 'non-existing',
       voice: 'non-existing',
     });
-    t.ok(purgedCountWhenErrored === 0, `purged no records when specified key was not found`);
-    t.ok(error, `error returned when specified key was not found`);
+    t.ok(purgedCountWhenErrored === 0, 'purged no records when specified key was not found');
+    t.ok(error, 'error returned when specified key was not found');
 
     // make sure other tts keys are still there
-    const cached = await client.keys('tts:*')
-    t.ok(cached.length >= 1, `successfully kept all non-specified tts records in cache`);
+    const cached = await client.keys('tts:*');
+    t.ok(cached.length >= 1, 'successfully kept all non-specified tts records in cache');
 
-    // retrieve keys from cache and check the key contains the file extension    
-    let key = cached[0];
-    t.ok(key.includes('mp3'), `tts cache extension shoult be part of the key and equal mp3`);
-
-    process.env.VG_TRIM_TTS_SILENCE = 'true';    
+    process.env.VG_TRIM_TTS_SILENCE = 'true';
     await client.set(makeSynthKey({ vendor: 'azure' }), 'value');
-
-    const r8Keys = await client.keys('tts:r8*');
-    key = r8Keys[0];
-    t.ok(key.includes('r8'), `tts cache extension shoult be part of the key and equal r8`);
-
   } catch (err) {
     console.error(JSON.stringify(err));
     t.end(err);