unSpeech TypeScript Client

Your Text-to-Speech Services, All-in-One.

Install

npm i unspeech

Getting Started

List voices

Besides of the /audio/speech endpoint, we support listing all the available voices from providers as well:

import { createUnSpeech, listVoices } from 'unspeech'

const unspeech = createUnSpeech('YOUR_EXTERNAL_PROVIDER_API_KEY', 'http://localhost:5933/v1/')

const voices = await listVoices(
  unspeech.voice({ backend: 'elevenlabs' })
)

Speech synthesis

For general purpose /audio/speech requests, @xsai/generate-speech or xsAI can be used as it's compatible:

npm i @xsai/generate-speech

import { generateSpeech } from '@xsai/generate-speech'
import { createUnSpeech } from 'unspeech'

const unspeech = createUnSpeech('YOUR_EXTERNAL_PROVIDER_API_KEY', 'http://localhost:5933/v1/')
const speech = await generateSpeech({
  ...unspeech.speech('elevenlabs/eleven_multilingual_v2'),
  input: 'Hello, World!',
  voice: '9BWtsMINqrJLrRacOk9x',
})

For the other providers, you can import them as needed

import {
  createUnAlibabaCloud,
  createUnElevenLabs,
  createUnMicrosoft,
  createUnSpeech,
  createUnVolcengine,
} from 'unspeech'

When using

• Microsoft / Azure AI Speech service
• Alibaba Cloud Model Studio / 阿里云百炼 / CosyVoice
• Volcano Engine / 火山引擎语音技术
• ElevenLabs

providers, SSML is supported to control in fine grain level for pitch, volume, rate, etc.

Related Projects

Looking for something like unSpeech, but for local TTS? check it out:

• erew123/alltalk_tts/alltalkbeta
• astramind-ai/Auralis
• matatonic/openedai-speech

Or to use free Edge TTS:

• travisvn/openai-edge-tts

License

AGPL-3.0