Video coming soon…
🤖 Setup Speaches — Self-Hosted Speech-to-Text & TTS API
Deploy Speaches on Ubuntu with Docker — an OpenAI-compatible speech-to-text and text-to-speech server powered by faster-whisper, a drop-in for the OpenAI audio API, on port 8192.
⚠️ This script is provided for demo and testing purposes only. Not intended for production use.
📦 Resources & Setup Scripts
Grab the automated bash script from GitHub to follow along with the video.
Quick Install:
wget https://raw.githubusercontent.com/mhmdali94/Docker/main/ai/speaches/speaches-ubuntu.sh
chmod +x speaches-ubuntu.sh
sudo bash speaches-ubuntu.shTutorial Steps
1 Download the Script
wget https://raw.githubusercontent.com/mhmdali94/Docker/main/ai/speaches/speaches-ubuntu.sh2 Make it Executable
chmod +x speaches-ubuntu.sh3 Run the Installer
The script installs Docker if needed and starts the CPU build of Speaches on port 8192, with model files cached in ./cache so they aren't re-downloaded each restart.
sudo bash speaches-ubuntu.sh4 Test the Transcription Endpoint
Speaches has no web UI — it's an API. Test transcription with a POST to the OpenAI-compatible endpoint:
curl -X POST http://<your-server-ip>:8192/v1/audio/transcriptionsPorts Used
| Port | Purpose |
|---|---|
| 8192 | OpenAI-compatible audio API (host → container 8000) — no web UI |
| ./cache | Downloaded model files |
| build | CPU build by default; GPU builds available for volume |