Remade with Rust
FFAI: on-device AI media library
The AI media library — ASR, TTS, OCR, VLM — one Rust surface for speech, text, images, and video
Replaces: Whisper.cpp / Tesseract / model glue
What is FFAI, the on-device AI media library?
FFAI is one Rust surface for speech, text, images, and video. It replaces the usual Whisper.cpp, Tesseract, and model-glue stack so ASR, TTS, OCR, and VLM stay in one library.
What FFAI replaces
FFAI is the on-device AI media library for speech, text, images, and video. ASR, TTS, OCR, and VLM share one Rust surface instead of a pile of model glue.
Why FFAI is a on-device AI media library
What FFAI keeps from Whisper.cpp / Tesseract / model glue
FFAI keeps the jobs those tools already do: hear speech, read text in an image, and look at a frame. Mercury and Carmenta are engines inside that surface, and each is also published on its own.
Where FFAI fits
FFAI is the AI side of the media stack next to ffmpeg. A product that needs speech or OCR starts here, then drops to Mercury or Carmenta when it wants one engine alone.
Read more about FFAI
- All Remade with Rust rebuilds
- Freedom Guide
- Vision
- Mercury — pure-Rust Whisper speech engine
- Carmenta — pure-Rust OCR engine
- Learn article on FFAI
- FFAI on GitHub