Speaches

• Source: https://github.com/speaches-ai/speaches
• License: MIT
• Alternatives: Chatterbox, Piper, Coqui TTS
• Status: TBD

Text-to-speech service.

Features

• OpenAI-Compatible API: Drop-in replacement for the OpenAI TTS endpoint
• Multiple Voice Models: Supports Kokoro, Piper, and other open-source TTS models
• Streaming Audio: Low-latency streaming output for real-time applications
• Multiple Formats: Output in MP3, WAV, FLAC, Opus, and AAC
• Self-Hosted & Private: No data sent to external services