Unisound

Unisound Token Hub API

Multi-modal MaaS API for Unisound U2 models: OpenAI-compatible chat completions and Anthropic-compatible messages for the U2 and U2-Med LLMs (streaming, function calling, structured output, context caching, prefix continuation), WebSocket one-shot and real-time speech recognition, asynchronous long-audio transcription and long-text speech synthesis via task polling, voiceprint registration and lookup, voice cloning and timbre management, asynchronous document parsing and information extraction (U1-OCR), and file management. Authenticated with Bearer API keys.

GitHub

Documentation

📖
Documentation
https://maas.unisound.com/docs
📖
APIReference
https://maas.unisound.com/docs/api/overview

Other Resources

🔗
APIsJSON
https://raw.githubusercontent.com/api-evangelist/unisound/refs/heads/main/apis.yml

API entry from apis.yml

apis.yml Raw ↑ 
name: Unisound Token Hub API
description: 'Multi-modal MaaS API for Unisound U2 models: OpenAI-compatible chat completions and Anthropic-compatible
  messages for the U2 and U2-Med LLMs (streaming, function calling, structured output, context caching,
  prefix continuation), WebSocket one-shot and real-time speech recognition, asynchronous long-audio transcription
  and long-text speech synthesis via task polling, voiceprint registration and lookup, voice cloning and
  timbre management, asynchronous document parsing and information extraction (U1-OCR), and file management.
  Authenticated with Bearer API keys.'
humanURL: https://maas.unisound.com/docs/api/overview
baseURL: https://maas-api.unisound.com/v1
tags:
- LLM
- Speech Recognition
- Text-to-Speech
- Voice Cloning
- OCR
properties:
- type: Documentation
  url: https://maas.unisound.com/docs
- type: APIReference
  url: https://maas.unisound.com/docs/api/overview