Speaches
ghcr.io/speaches-ai/speaches:latest-cpu
ghcr.io/speaches-ai/speaches
latest-cpu
Latest stable release
latest-cuda
Latest stable release with CUDA support
bridge
http://[IP]:[PORT:8000]/
false
https://speaches.ai/configuration/
https://speaches.ai/
speaches is an OpenAI API-compatible server supporting streaming transcription, translation, and speech generation. Speach-to-Text is powered by faster-whisper and for Text-to-Speech piper and Kokoro are used. This project aims to be Ollama, but for TTS/STT models.
[br]
**Nvidia GPU Use:**
Using the Unraid Nvidia Plugin to install a version of Unraid with the Nvidia Drivers installed and add **--runtime=nvidia --gpus=all** to [b]"extra parameters"[/b] (switch on advanced view)
False
AI: Productivity: Tools: Other: Status:Stable
whisper speech text voice generator piper faster openai gpu cuda nvidia tts stt
https://raw.githubusercontent.com/nwithan8/unraid_templates/master/images/speaches-icon.png
https://raw.githubusercontent.com/nwithan8/unraid_templates/main/templates/speaches.xml
https://github.com/nwithan8
### 2025-04-16
Initial release
8000
/mnt/user/appdata/speaches/cache