0.0.42

support whisper streaming

.github/workflows/ci.yml

@@ -14,6 +14,9 @@ jobs:
           node-version: lts/*
       - run: npm install
       - run: npm run jslint
+      - run: sudo apt update && sudo apt install -y squid
+      - run: sudo cp test/squid.conf /etc/squid/squid.conf
+      - run: sudo systemctl start squid
       - run: npm test
         env:
           AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_ID }}
@@ -24,4 +27,10 @@ jobs:
           IBM_TTS_API_KEY: ${{ secrets.IBM_TTS_API_KEY }}
           IBM_TTS_REGION: ${{ secrets.IBM_TTS_REGION }}
           MICROSOFT_API_KEY: ${{ secrets.MICROSOFT_API_KEY }}
-          MICROSOFT_REGION: ${{ secrets.MICROSOFT_REGION }}
+          MICROSOFT_REGION: ${{ secrets.MICROSOFT_REGION }}
+          OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
+          ELEVENLABS_API_KEY: ${{ secrets.ELEVENLABS_API_KEY }}
+          ELEVENLABS_VOICE_ID: ${{ secrets.ELEVENLABS_VOICE_ID }}
+          ELEVENLABS_MODEL_ID: ${{ secrets.ELEVENLABS_MODEL_ID }}
+          JAMBONES_HTTP_PROXY_IP: 127.0.0.1
+          JAMBONES_HTTP_PROXY_PORT: 3128

README.md

@@ -1,2 +1,3 @@
-# speech-utils
-TTS-related speech utilities for jambonz
+# speech-utils ![CI](https://github.com/jambonz/speech-utils/workflows/CI/badge.svg)
+TTS-related speech utilities for jambonz.
+

index.js

@@ -1,41 +1,22 @@
 const {noopLogger} = require('./lib/utils');
-const Redis = require('ioredis');
 
 module.exports = (opts, logger) => {
   logger = logger || noopLogger;
-  const connectionOpts = {...opts};
-  // Support legacy app
-  if (process.env.JAMBONES_REDIS_USERNAME && process.env.JAMBONES_REDIS_PASSWORD) {
-    if (Array.isArray(connectionOpts)) {
-      for (const o of opts) {
-        o.username = process.env.JAMBONES_REDIS_USERNAME;
-        o.password = process.env.JAMBONES_REDIS_PASSWORD;
-      }
-    } else {
-      connectionOpts.username = process.env.JAMBONES_REDIS_USERNAME;
-      connectionOpts.password = process.env.JAMBONES_REDIS_PASSWORD;
-    }
-  }
-
-  const client = new Redis(connectionOpts);
-  ['ready', 'connect', 'reconnecting', 'error', 'end', 'warning']
-    .forEach((event) => {
-      client.on(event, (...args) => {
-        if ('error' === event) {
-          if (process.env.NODE_ENV === 'test' && args[0]?.code === 'ECONNREFUSED') return;
-          logger.error({...args}, '@jambonz/realtimedb-helpers - redis error');
-        }
-        else logger.debug({args}, `redis event ${event}`);
-      });
-    });
+  const {
+    client,
+    createHash,
+    retrieveHash
+  } = require('@jambonz/realtimedb-helpers')(opts, logger);
 
   return {
     client,
     getTtsSize: require('./lib/get-tts-size').bind(null, client, logger),
     purgeTtsCache: require('./lib/purge-tts-cache').bind(null, client, logger),
+    addFileToCache: require('./lib/add-file-to-cache').bind(null, client, logger),
     synthAudio: require('./lib/synth-audio').bind(null, client, logger),
     getNuanceAccessToken: require('./lib/get-nuance-access-token').bind(null, client, logger),
     getIbmAccessToken: require('./lib/get-ibm-access-token').bind(null, client, logger),
+    getAwsAuthToken: require('./lib/get-aws-sts-token').bind(null, logger, createHash, retrieveHash),
     getTtsVoices: require('./lib/get-tts-voices').bind(null, client, logger),
   };
 };

lib/add-file-to-cache.js

@@ -0,0 +1,30 @@
+const fs = require('fs/promises');
+const {noopLogger, makeSynthKey} = require('./utils');
+const EXPIRES = (process.env.JAMBONES_TTS_CACHE_DURATION_MINS || 4 * 60) * 60; // cache tts for 4 hours
+
+async function addFileToCache(client, logger, path,
+  {account_sid, vendor, language, voice, deploymentId, engine, text}) {
+  let key;
+  logger = logger || noopLogger;
+
+  try {
+    key = makeSynthKey({
+      account_sid,
+      vendor,
+      language: language || '',
+      voice: voice || deploymentId,
+      engine,
+      text,
+    });
+    const audioBuffer = await fs.readFile(path);
+    await client.setex(key, EXPIRES, audioBuffer.toString('base64'));
+  } catch (err) {
+    logger.error(err, 'addFileToCache: Error');
+    return;
+  }
+
+  logger.debug(`addFileToCache: added ${path} to cache with key ${key}`);
+  return key;
+}
+
+module.exports = addFileToCache;

lib/get-aws-sts-token.js

@@ -0,0 +1,45 @@
+const { STSClient, GetSessionTokenCommand } = require('@aws-sdk/client-sts');
+const {makeAwsKey, noopLogger} = require('./utils');
+const debug = require('debug')('jambonz:speech-utils');
+const EXPIRY = 3600;
+
+async function getAwsAuthToken(
+  logger,
+  createHash, retrieveHash,
+  awsAccessKeyId, awsSecretAccessKey, awsRegion) {
+  logger = logger || noopLogger;
+  try {
+    const key = makeAwsKey(awsAccessKeyId);
+    const obj = await retrieveHash(key);
+    if (obj) return {...obj, servedFromCache: true};
+
+    /* access token not found in cache, so generate it using STS */
+    const stsClient = new STSClient({
+      region: awsRegion,
+      credentials: {
+        accessKeyId: awsAccessKeyId,
+        secretAccessKey: awsSecretAccessKey,
+      }
+    });
+    const command = new GetSessionTokenCommand({DurationSeconds: EXPIRY});
+    const data = await stsClient.send(command);
+
+    const credentials = {
+      accessKeyId: data.Credentials.AccessKeyId,
+      secretAccessKey: data.Credentials.SecretAccessKey,
+      securityToken: data.Credentials.SessionToken
+    };
+
+    /* expire 10 minutes before the hour, so we don't lose the use of it during a call */
+    createHash(key, credentials, EXPIRY - 600)
+      .catch((err) => logger.error(err, `Error saving hash for key ${key}`));
+
+    return {...credentials, servedFromCache: false};
+  } catch (err) {
+    debug(err, 'getAwsAuthToken: Error retrieving AWS auth token');
+    logger.error(err, 'getAwsAuthToken: Error retrieving AWS auth token');
+    throw err;
+  }
+}
+
+module.exports = getAwsAuthToken;

lib/get-tts-voices.js

@@ -97,7 +97,7 @@ const getAwsVoices = async(_client, logger, credentials) => {
         secretAccessKey
       }
     });
-    const command = new DescribeVoicesCommand({LanguageCode: 'en-US'});
+    const command = new DescribeVoicesCommand({});
     const response = await client.send(command);
     return response;
   } catch (err) {

lib/synth-audio.js

@@ -14,7 +14,13 @@ const {
   CancellationDetails,
   SpeechSynthesisOutputFormat
 } = sdk;
-const {makeSynthKey, createNuanceClient, createKryptonClient, createRivaClient, noopLogger} = require('./utils');
+const {
+  makeSynthKey,
+  createNuanceClient,
+  createKryptonClient,
+  createRivaClient,
+  noopLogger
+} = require('./utils');
 const getNuanceAccessToken = require('./get-nuance-access-token');
 const {
   SynthesisRequest,
@@ -32,6 +38,7 @@ const {AudioEncoding} = require('../stubs/riva/proto/riva_audio_pb');
 const debug = require('debug')('jambonz:realtimedb-helpers');
 const EXPIRES = (process.env.JAMBONES_TTS_CACHE_DURATION_MINS || 4 * 60) * 60; // cache tts for 4 hours
 const TMP_FOLDER = '/tmp';
+const OpenAI = require('openai');
 
 
 const trimTrailingSilence = (buffer) => {
@@ -69,14 +76,16 @@ const trimTrailingSilence = (buffer) => {
  * the synthesized audio, and a variable indicating whether it was served from cache
  */
 async function synthAudio(client, logger, stats, { account_sid,
-  vendor, language, voice, gender, text, engine, salt, model, credentials, deploymentId, disableTtsCache
+  vendor, language, voice, gender, text, engine, salt, model, credentials, deploymentId,
+  disableTtsCache, renderForCaching, options
 }) {
   let audioBuffer;
   let servedFromCache = false;
   let rtt;
   logger = logger || noopLogger;
 
-  assert.ok(['google', 'aws', 'polly', 'microsoft', 'wellsaid', 'nuance', 'nvidia', 'ibm'].includes(vendor) ||
+  assert.ok(['google', 'aws', 'polly', 'microsoft',
+    'wellsaid', 'nuance', 'nvidia', 'ibm', 'elevenlabs', 'whisper', 'deepgram'].includes(vendor) ||
   vendor.startsWith('custom'),
   `synthAudio supported vendors are google, aws, microsoft, nuance, nvidia and wellsaid, not ${vendor}`);
   if ('google' === vendor) {
@@ -110,6 +119,14 @@ async function synthAudio(client, logger, stats, { account_sid,
     language = 'en-US'; // WellSaid only supports English atm
     assert.ok(voice, 'synthAudio requires voice when wellsaid is used');
     assert.ok(!text.startsWith('<speak'), 'wellsaid does not support SSML tags');
+  } else if ('elevenlabs' === vendor) {
+    assert.ok(voice, 'synthAudio requires voice when elevenlabs is used');
+    assert.ok(credentials.api_key, 'synthAudio requires api_key when elevenlabs is used');
+    assert.ok(credentials.model_id, 'synthAudio requires model_id when elevenlabs is used');
+  } else if ('whisper' === vendor) {
+    assert.ok(voice, 'synthAudio requires voice when whisper is used');
+    assert.ok(credentials.model_id, 'synthAudio requires model when whisper is used');
+    assert.ok(credentials.api_key, 'synthAudio requires api_key when whisper is used');
   } else  if (vendor.startsWith('custom')) {
     assert.ok(credentials.custom_tts_url, `synthAudio requires custom_tts_url in credentials when ${vendor} is used`);
   }
@@ -125,8 +142,11 @@ async function synthAudio(client, logger, stats, { account_sid,
   if (['nuance', 'nvidia'].includes(vendor) ||
     (
       process.env.JAMBONES_TTS_TRIM_SILENCE &&
-      ['microsoft', 'azure'].includes(vendor) &&
-      !text.startsWith('<speak')
+      ['microsoft', 'azure'].includes(vendor)
+    ) ||
+    (
+      !process.env.JAMBONES_DISABLE_TTS_STREAMING &&
+      vendor === 'elevenlabs'
     )
   ) {
     filePath = `${TMP_FOLDER}/${key.replace('tts:', `tts-${salt || ''}`)}.r8`;
@@ -178,6 +198,27 @@ async function synthAudio(client, logger, stats, { account_sid,
       case 'wellsaid':
         audioBuffer = await synthWellSaid(logger, {credentials, stats, language, voice, text, filePath});
         break;
+      case 'elevenlabs':
+        audioBuffer = await synthElevenlabs(logger, {
+          credentials, options, stats, language, voice, text, renderForCaching, filePath
+        });
+        if (typeof audioBuffer === 'object' && audioBuffer.filePath) {
+          return audioBuffer;
+        }
+        else {
+          audioBuffer = await synthElevenlabs(logger, {credentials, options, stats, language, voice, text, filePath});
+        }
+        break;
+      case 'whisper':
+        audioBuffer = await synthWhisper(logger, {credentials, stats, voice, text, renderForCaching});
+        if (typeof audioBuffer === 'object' && audioBuffer.filePath) {
+          return audioBuffer;
+        }
+        audioBuffer = await synthWhisper(logger, {credentials, stats, voice, text});
+        break;
+      case 'deepgram':
+        audioBuffer = await synthDeepgram(logger, {credentials, stats, model, text});
+        break;
       case vendor.startsWith('custom') ? vendor : 'cant_match_value':
         ({ audioBuffer, filePath } = await synthCustomVendor(logger,
           {credentials, stats, language, voice, text, filePath}));
@@ -248,7 +289,8 @@ const synthGoogle = async(logger, {credentials, stats, language, voice, gender,
   const client = new ttsGoogle.TextToSpeechClient(credentials);
   const opts = {
     voice: {
-      name: voice,
+      ...(typeof voice === 'string' && {name: voice}),
+      ...(typeof voice === 'object' && {customVoice: voice}),
       languageCode: language,
       ssmlGender: gender || 'SSML_VOICE_GENDER_UNSPECIFIED'
     },
@@ -298,6 +340,48 @@ const synthIbm = async(logger, {credentials, stats, voice, text}) => {
   }
 };
 
+async function _synthOnPremMicrosoft(logger, {
+  credentials,
+  stats,
+  language,
+  voice,
+  text,
+  filePath
+}) {
+  const {use_custom_tts, custom_tts_endpoint_url} = credentials;
+  let content = text;
+
+  if (use_custom_tts && !content.startsWith('<speak')) {
+    /**
+     * Note: it seems that to use custom voice ssml is required with the voice attribute
+     * Otherwise sending plain text we get "Voice does not match"
+     */
+    content = `<speak>${text}</speak>`;
+  }
+
+  if (content.startsWith('<speak>')) {
+    /* microsoft enforces some properties and uses voice xml element so if the user did not supply do it for them */
+    const words = content.slice(7, -8).trim().replace(/(\r\n|\n|\r)/gm, ' ');
+    // eslint-disable-next-line max-len
+    content = `<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xmlns:mstts="https://www.w3.org/2001/mstts" xml:lang="${language}"><voice name="${voice}">${words}</voice></speak>`;
+    logger.info({content}, 'synthMicrosoft');
+  }
+
+  try {
+    const trimSilence = filePath.endsWith('.r8');
+    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'
+    });
+    const mp3 = await post(custom_tts_endpoint_url, content);
+    return mp3;
+  } catch (err) {
+    logger.info({err}, '_synthMicrosoftByHttp returned error');
+    throw err;
+  }
+}
+
 const synthMicrosoft = async(logger, {
   credentials,
   stats,
@@ -307,7 +391,17 @@ const synthMicrosoft = async(logger, {
   filePath
 }) => {
   try {
-    const {api_key: apiKey, region, use_custom_tts, custom_tts_endpoint} = credentials;
+    const {api_key: apiKey, region, use_custom_tts, custom_tts_endpoint, custom_tts_endpoint_url} = credentials;
+    if (use_custom_tts && custom_tts_endpoint_url) {
+      return await _synthOnPremMicrosoft(logger, {
+        credentials,
+        stats,
+        language,
+        voice,
+        text,
+        filePath
+      });
+    }
     const trimSilence = filePath.endsWith('.r8');
     let content = text;
     const speechConfig = SpeechConfig.fromSubscription(apiKey, region);
@@ -315,16 +409,23 @@ const synthMicrosoft = async(logger, {
     speechConfig.speechSynthesisVoiceName = voice;
     if (use_custom_tts && custom_tts_endpoint) {
       speechConfig.endpointId = custom_tts_endpoint;
-
+    }
+    if (use_custom_tts && !content.startsWith('<speak')) {
       /**
        * Note: it seems that to use custom voice ssml is required with the voice attribute
        * Otherwise sending plain text we get "Voice does not match"
        */
-      if (!content.startsWith('<speak')) content = `<speak>${text}</speak>`;
+      content = `<speak>${text}</speak>`;
     }
     speechConfig.speechSynthesisOutputFormat = trimSilence ?
       SpeechSynthesisOutputFormat.Raw8Khz16BitMonoPcm :
       SpeechSynthesisOutputFormat.Audio16Khz32KBitRateMonoMp3;
+
+    if (process.env.JAMBONES_HTTP_PROXY_IP && process.env.JAMBONES_HTTP_PROXY_PORT) {
+      logger.debug(
+        `synthMicrosoft: using proxy ${process.env.JAMBONES_HTTP_PROXY_IP}:${process.env.JAMBONES_HTTP_PROXY_PORT}`);
+      speechConfig.setProxy(process.env.JAMBONES_HTTP_PROXY_IP, process.env.JAMBONES_HTTP_PROXY_PORT);
+    }
     const synthesizer = new SpeechSynthesizer(speechConfig);
 
     if (content.startsWith('<speak>')) {
@@ -510,6 +611,113 @@ const synthCustomVendor = async(logger, {credentials, stats, language, voice, te
   }
 };
 
+const synthElevenlabs = async(logger, {credentials, options, stats, language, voice, text, renderForCaching}) => {
+  const {api_key, model_id, options: credOpts} = credentials;
+  const opts = !!options && Object.keys(options).length !== 0 ? options : JSON.parse(credOpts || '{}');
+
+  /* default to using the streaming interface, unless disabled by env var OR we want just a cache file */
+  if (!process.env.JAMBONES_DISABLE_TTS_STREAMING && !renderForCaching) {
+    let params = '';
+    params += `{api_key=${api_key}`;
+    params += `,model_id=${model_id}`;
+    params += `,optimize_streaming_latency=${opts.optimize_streaming_latency || 2}`;
+    params += ',write_cache_file=1';
+    if (opts.voice_settings?.similarity_boost) params += `,similarity_boost=${opts.voice_settings.similarity_boost}`;
+    if (opts.voice_settings?.stability) params += `,stability=${opts.voice_settings.stability}`;
+    if (opts.voice_settings?.style) params += `,style=${opts.voice_settings.style}`;
+    if (opts.voice_settings?.use_speaker_boost === false) params += ',use_speaker_boost=false';
+    params += '}';
+
+    return {
+      filePath: `say:${params}${text.replace(/\n/g, ' ').replace(/\r/g, ' ')}`,
+      servedFromCache: false,
+      rtt: 0
+    };
+  }
+
+  const optimize_streaming_latency = opts.optimize_streaming_latency ?
+    `?optimize_streaming_latency=${opts.optimize_streaming_latency}` : '';
+  try {
+    const post = bent('https://api.elevenlabs.io', 'POST', 'buffer', {
+      'xi-api-key': api_key,
+      'Accept': 'audio/mpeg',
+      'Content-Type': 'application/json'
+    });
+    const mp3 = await post(`/v1/text-to-speech/${voice}${optimize_streaming_latency}`, {
+      text,
+      model_id,
+      voice_settings: {
+        stability: 0.5,
+        similarity_boost: 0.5
+      },
+      ...opts
+    });
+    return mp3;
+  } catch (err) {
+    logger.info({err}, 'synth Elevenlabs returned error');
+    stats.increment('tts.count', ['vendor:elevenlabs', 'accepted:no']);
+    throw err;
+  }
+};
+
+const synthWhisper = async(logger, {credentials, stats, voice, text, renderForCaching}) => {
+  const {api_key, model_id, baseURL, timeout, speed} = credentials;
+  /* if the env is set to stream then bag out, unless we are specifically rendering to generate a cache file */
+  if (!process.env.JAMBONES_DISABLE_TTS_STREAMING && !renderForCaching) {
+    let params = '';
+    params += `{api_key=${api_key}`;
+    params += `,model_id=${model_id}`;
+    params += `,voice=${voice}`;
+    params += ',write_cache_file=1';
+    if (speed) params += `,speed=${speed}`;
+    params += '}';
+
+    return {
+      filePath: `say:${params}${text.replace(/\n/g, ' ')}`,
+      servedFromCache: false,
+      rtt: 0
+    };
+  }
+  try {
+    const openai = new OpenAI.OpenAI({
+      apiKey: api_key,
+      timeout: timeout || 5000,
+      ...(baseURL && {baseURL})
+    });
+
+    const mp3 = await openai.audio.speech.create({
+      model: model_id,
+      voice,
+      input: text,
+      response_format: 'mp3'
+    });
+    return Buffer.from(await mp3.arrayBuffer());
+  } catch (err) {
+    logger.info({err}, 'synth whisper returned error');
+    stats.increment('tts.count', ['vendor:openai', 'accepted:no']);
+    throw err;
+  }
+};
+
+const synthDeepgram = async(logger, {credentials, stats, model, text}) => {
+  const {api_key} = credentials;
+  try {
+    const post = bent('https://api.beta.deepgram.com', 'POST', 'buffer', {
+      'Authorization': `Token ${api_key}`,
+      'Accept': 'audio/mpeg',
+      'Content-Type': 'application/json'
+    });
+    const mp3 = await post(`/v1/speak?model=${model}`, {
+      text
+    });
+    return mp3;
+  } catch (err) {
+    logger.info({err}, 'synth Deepgram returned error');
+    stats.increment('tts.count', ['vendor:deepgram', 'accepted:no']);
+    throw err;
+  }
+};
+
 const getFileExtFromMime = (mime) => {
   switch (mime) {
     case 'audio/wav':

lib/utils.js

@@ -43,6 +43,12 @@ function makeIbmKey(apiKey) {
   return `ibm:${hash.digest('hex')}`;
 }
 
+function makeAwsKey(awsAccessKeyId) {
+  const hash = crypto.createHash('sha1');
+  hash.update(awsAccessKeyId);
+  return `aws:${hash.digest('hex')}`;
+}
+
 function makeNuanceKey(clientId, secret, scope) {
   const hash = crypto.createHash('sha1');
   hash.update(`${clientId}:${secret}:${scope}`);
@@ -106,11 +112,11 @@ const createRivaClient = async(rivaUri) => {
   return client;
 };
 
-
 module.exports = {
   makeSynthKey,
   makeNuanceKey,
   makeIbmKey,
+  makeAwsKey,
   getNuanceAccessToken,
   createNuanceClient,
   createKryptonClient,

package.json

@@ -1,6 +1,6 @@
 {
   "name": "@jambonz/speech-utils",
-  "version": "0.0.18",
+  "version": "0.0.42",
   "description": "TTS-related speech utilities for jambonz",
   "main": "index.js",
   "author": "Dave Horton",
@@ -24,24 +24,26 @@
   },
   "homepage": "https://github.com/jambonz/speech-utils#readme",
   "dependencies": {
-    "@aws-sdk/client-polly": "^3.359.0",
-    "@google-cloud/text-to-speech": "^4.2.1",
-    "@grpc/grpc-js": "^1.8.13",
+    "@aws-sdk/client-polly": "^3.496.0",
+    "@aws-sdk/client-sts": "^3.496.0",
+    "@google-cloud/text-to-speech": "^5.0.2",
+    "@grpc/grpc-js": "^1.9.14",
+    "@jambonz/realtimedb-helpers": "^0.8.7",
     "bent": "^7.3.12",
     "debug": "^4.3.4",
-    "form-urlencoded": "^6.1.0",
+    "form-urlencoded": "^6.1.4",
     "google-protobuf": "^3.21.2",
     "ibm-watson": "^8.0.0",
-    "microsoft-cognitiveservices-speech-sdk": "^1.26.0",
-    "ioredis": "^5.3.2",
-    "undici": "^5.21.0"
+    "microsoft-cognitiveservices-speech-sdk": "1.34.0",
+    "openai": "^4.25.0",
+    "undici": "^6.4.0"
   },
   "devDependencies": {
-    "config": "^3.3.9",
-    "eslint": "^8.33.0",
+    "config": "^3.3.10",
+    "eslint": "^8.56.0",
     "eslint-plugin-promise": "^6.1.1",
     "nyc": "^15.1.0",
-    "pino": "^7.2.0",
-    "tape": "^5.1.1"
+    "pino": "^8.17.0",
+    "tape": "^5.7.3"
   }
 }

test/aws.js

@@ -0,0 +1,40 @@
+const test = require('tape').test ;
+const config = require('config');
+const opts = config.get('redis');
+const logger = require('pino')({level: 'error'});
+process.on('unhandledRejection', (reason, p) => {
+  console.log('Unhandled Rejection at: Promise', p, 'reason:', reason);
+});
+
+const sleep = (ms) => new Promise((resolve) => setTimeout(resolve, ms));
+
+test('AWS - create and cache auth token', async(t) => {
+  const fn = require('..');
+  const {client, getAwsAuthToken} = fn(opts, logger);
+
+  if (!process.env.AWS_ACCESS_KEY_ID || !process.env.AWS_SECRET_ACCESS_KEY || !process.env.AWS_REGION) {
+      t.pass('skipping AWS auth token tests since no AWS credentials provided');
+      t.end();
+      client.quit();
+      return;
+  }
+  try {
+    let obj = await getAwsAuthToken(process.env.AWS_ACCESS_KEY_ID, process.env.AWS_SECRET_ACCESS_KEY, process.env.AWS_REGION);
+    //console.log({obj}, 'received auth token from AWS');
+    t.ok(obj.securityToken && !obj.servedFromCache, 'successfullY generated auth token from AWS');
+
+    await sleep(250);
+    obj = await getAwsAuthToken(process.env.AWS_ACCESS_KEY_ID, process.env.AWS_SECRET_ACCESS_KEY, process.env.AWS_REGION);
+    //console.log({obj}, 'received auth token from AWS - second request');
+    t.ok(obj.securityToken && obj.servedFromCache, 'successfully received access token from cache');
+ 
+    await client.flushall();
+    t.end();
+  }
+  catch (err) {
+    console.error(err);
+    t.end(err);
+  }
+  client.quit();
+});
+

test/index.js

@@ -1,4 +1,7 @@
 require('./docker_start');
 require('./synth');
 require('./list-voices');
+require('./aws');
+require('./ibm');
+require('./nuance');
 require('./docker_stop');

test/squid.conf

@@ -0,0 +1,85 @@
+#
+# Recommended minimum configuration:
+#
+
+# Example rule allowing access from your local networks.
+# Adapt to list your (internal) IP networks from where browsing
+# should be allowed
+acl localnet src 0.0.0.1-0.255.255.255	# RFC 1122 "this" network (LAN)
+acl localnet src 10.0.0.0/8		# RFC 1918 local private network (LAN)
+acl localnet src 100.64.0.0/10		# RFC 6598 shared address space (CGN)
+acl localnet src 169.254.0.0/16 	# RFC 3927 link-local (directly plugged) machines
+acl localnet src 172.16.0.0/12		# RFC 1918 local private network (LAN)
+acl localnet src 192.168.0.0/16		# RFC 1918 local private network (LAN)
+acl localnet src fc00::/7       	# RFC 4193 local private network range
+acl localnet src fe80::/10      	# RFC 4291 link-local (directly plugged) machines
+
+acl SSL_ports port 443
+acl Safe_ports port 80		# http
+acl Safe_ports port 21		# ftp
+acl Safe_ports port 443		# https
+acl Safe_ports port 70		# gopher
+acl Safe_ports port 210		# wais
+acl Safe_ports port 1025-65535	# unregistered ports
+acl Safe_ports port 280		# http-mgmt
+acl Safe_ports port 488		# gss-http
+acl Safe_ports port 591		# filemaker
+acl Safe_ports port 777		# multiling http
+
+#
+# Recommended minimum Access Permission configuration:
+#
+# Deny requests to certain unsafe ports
+http_access deny !Safe_ports
+
+# Deny CONNECT to other than secure SSL ports
+http_access allow CONNECT !SSL_ports
+
+# Only allow cachemgr access from localhost
+http_access allow localhost manager
+http_access deny manager
+
+# This default configuration only allows localhost requests because a more
+# permissive Squid installation could introduce new attack vectors into the
+# network by proxying external TCP connections to unprotected services.
+http_access allow localhost
+
+# The two deny rules below are unnecessary in this default configuration
+# because they are followed by a "deny all" rule. However, they may become
+# critically important when you start allowing external requests below them.
+
+# Protect web applications running on the same server as Squid. They often
+# assume that only local users can access them at "localhost" ports.
+http_access deny to_localhost
+
+# Protect cloud servers that provide local users with sensitive info about
+# their server via certain well-known link-local (a.k.a. APIPA) addresses.
+http_access deny to_linklocal
+
+#
+# INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
+#
+
+# For example, to allow access from your local networks, you may uncomment the
+# following rule (and/or add rules that match your definition of "local"):
+# http_access allow localnet
+
+# And finally deny all other access to this proxy
+http_access deny all
+
+# Squid normally listens to port 3128
+http_port 3128
+
+# Uncomment and adjust the following to add a disk cache directory.
+#cache_dir ufs /usr/local/var/cache/squid 100 16 256
+
+# Leave coredumps in the first cache dir
+coredump_dir /usr/local/var/cache/squid
+
+#
+# Add any of your own refresh_pattern entries above these.
+#
+refresh_pattern ^ftp:		1440	20%	10080
+refresh_pattern ^gopher:	1440	0%	1440
+refresh_pattern -i (/cgi-bin/|\?) 0	0%	0
+refresh_pattern .		0	20%	4320

test/synth.js

@@ -20,7 +20,7 @@ const stats = {
 
 test('Google speech synth tests', async(t) => {
   const fn = require('..');
-  const {synthAudio, client} = fn(opts, logger);
+  const {synthAudio, addFileToCache, client} = fn(opts, logger);
 
   if (!process.env.GCP_FILE && !process.env.GCP_JSON_KEY) {
     t.pass('skipping google speech synth tests since neither GCP_FILE nor GCP_JSON_KEY provided');
@@ -38,7 +38,7 @@ test('Google speech synth tests', async(t) => {
         },
       },
       language: 'en-GB',
-      gender: 'MALE',
+      gender: 'FEMALE',
       text: 'This is a test.  This is only a test',
       salt: 'foo.bar',
     });
@@ -53,11 +53,19 @@ test('Google speech synth tests', async(t) => {
         },
       },
       language: 'en-GB',
-      gender: 'MALE',
+      gender: 'FEMALE',
       text: 'This is a test.  This is only a test',
     });
     t.ok(opts.servedFromCache, `successfully retrieved cached google audio from ${opts.filePath}`);
 
+    const success = await addFileToCache(opts.filePath, {
+      vendor: 'google',
+      language: 'en-GB',
+      gender: 'FEMALE',
+      text: 'This is a test.  This is only a test'
+    });
+    t.ok(success, `successfully added ${opts.filePath} to cache`);
+
     opts = await synthAudio(stats, {
       vendor: 'google',
       credentials: {
@@ -68,7 +76,7 @@ test('Google speech synth tests', async(t) => {
       },
       disableTtsCache: true,
       language: 'en-GB',
-      gender: 'MALE',
+      gender: 'FEMALE',
       text: 'This is a test.  This is only a test',
     });
     t.ok(!opts.servedFromCache, `successfully synthesized google audio regardless of current cache to ${opts.filePath}`);
@@ -79,6 +87,40 @@ test('Google speech synth tests', async(t) => {
   client.quit();
 });
 
+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');
+    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, {
+      vendor: 'google',
+      credentials: {
+        credentials: {
+          client_email: creds.client_email,
+          private_key: creds.private_key,
+        },
+      },
+      language: 'en-AU',
+      text: 'This is a test.  This is only a test',
+      voice: {
+        reportedUsage:"REALTIME",
+        model: process.env.GCP_CUSTOM_VOICE_MODEL
+      }
+    });
+    t.ok(!opts.servedFromCache, `successfully synthesized google custom voice 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);
@@ -148,7 +190,9 @@ test('Azure speech synth tests', async(t) => {
       text: longText,
     });
     t.ok(!opts.servedFromCache, `successfully synthesized microsoft audio to ${opts.filePath}`);
-
+    if (process.env.JAMBONES_HTTP_PROXY_IP && process.env.JAMBONES_HTTP_PROXY_PORT) {
+      t.pass('successfully used proxy to reach microsoft tts service');
+    }
 
     opts = await synthAudio(stats, {
       vendor: 'microsoft',
@@ -411,6 +455,100 @@ test('Custom Vendor speech synth tests', async(t) => {
   client.quit();
 });
 
+test('Elevenlabs speech synth tests', async(t) => {
+  const fn = require('..');
+  const {synthAudio, client} = fn(opts, logger);
+
+  if (!process.env.ELEVENLABS_API_KEY || !process.env.ELEVENLABS_VOICE_ID || !process.env.ELEVENLABS_MODEL_ID) {
+    t.pass('skipping ElevenLabs speech synth tests since ELEVENLABS_API_KEY or ELEVENLABS_VOICE_ID or ELEVENLABS_MODEL_ID not provided');
+    return t.end();
+  }
+  const text = 'Hi there and welcome to jambones!';
+  try {
+    let opts = await synthAudio(stats, {
+      vendor: 'elevenlabs',
+      credentials: {
+        api_key: process.env.ELEVENLABS_API_KEY,
+        model_id: process.env.ELEVENLABS_MODEL_ID,
+        options: JSON.stringify({
+          optimize_streaming_latency: 1,
+          voice_settings: {
+            similarity_boost: 1,
+            stability: 0.8,
+            style: 1,
+            use_speaker_boost: true
+          }
+        })
+      },
+      language: 'en-US',
+      voice: process.env.ELEVENLABS_VOICE_ID,
+      text,
+    });
+    t.ok(!opts.servedFromCache, `successfully synthesized eleven audio to ${opts.filePath}`);
+
+  } catch (err) {
+    console.error(JSON.stringify(err));
+    t.end(err);
+  }
+  client.quit();
+})
+
+test('whisper speech synth tests', async(t) => {
+  const fn = require('..');
+  const {synthAudio, client} = fn(opts, logger);
+
+  if (!process.env.OPENAI_API_KEY) {
+    t.pass('skipping OPENAI speech synth tests since OPENAI_API_KEY not provided');
+    return t.end();
+  }
+  const text = 'Hi there and welcome to jambones!';
+  try {
+    let opts = await synthAudio(stats, {
+      vendor: 'whisper',
+      credentials: {
+        api_key: process.env.OPENAI_API_KEY,
+        model_id: 'tts-1'
+      },
+      language: 'en-US',
+      voice: 'alloy',
+      text,
+    });
+    t.ok(!opts.servedFromCache, `successfully synthesized whisper audio to ${opts.filePath}`);
+
+  } catch (err) {
+    console.error(JSON.stringify(err));
+    t.end(err);
+  }
+  client.quit();
+})
+
+test('Deepgram speech synth tests', async(t) => {
+  const fn = require('..');
+  const {synthAudio, client} = fn(opts, logger);
+
+  if (!process.env.DEEPGRAM_API_KEY) {
+    t.pass('skipping Deepgram speech synth tests since DEEPGRAM_API_KEY');
+    return t.end();
+  }
+  const text = 'Hi there and welcome to jambones!';
+  try {
+    let opts = await synthAudio(stats, {
+      vendor: 'deepgram',
+      credentials: {
+        api_key: process.env.DEEPGRAM_API_KEY
+      },
+      model: 'alpha-asteria-en-v2',
+      text,
+    });
+    t.ok(!opts.servedFromCache, `successfully synthesized deepgram audio to ${opts.filePath}`);
+
+  } catch (err) {
+    console.error(JSON.stringify(err));
+    t.end(err);
+  }
+  client.quit();
+})
+
 test('TTS Cache tests', async(t) => {
   const fn = require('..');
   const {purgeTtsCache, getTtsSize, client} = fn(opts, logger);