# Soniox

> Soniox is a speech AI platform with real-time and async speech-to-text, text-to-speech, and translation APIs, built for the messy, multilingual, multi-speaker audio of the real world. It combines high accuracy across 60+ languages, code-switching, speaker diarization, reliable transcription of names and alphanumerics, and sub-200ms streaming latency, so developers can build voice agents, live translation, dictation, and call automation that work in production.

## Documentation
- [Docs overview](https://soniox.com/docs): Full API documentation
- [Docs structure overview](https://soniox.com/docs/llms.txt): Complete structure of docs in one file
- [API reference](https://soniox.com/docs/api-reference): REST and WebSocket APIs

## Products
- [Speech-to-Text](https://soniox.com/speech-to-text): Real-time and async transcription
- [Text-to-Speech](https://soniox.com/text-to-speech): Real-time and async voice generation
- [Translation](https://soniox.com/speech-translation): Real-time translation, already built into speech-to-text API
- [Pricing](https://soniox.com/pricing): API pricing

## Optional
- [Compare](https://soniox.com/compare): Soniox vs other providers
- [Blog](https://soniox.com/blog): Product and engineering posts