Dicklesworthstone

Dicklesworthstone / franken_whisper

Public

Agent-first Rust ASR orchestration stack: Bayesian backend routing across whisper.cpp/insanely-fast-whisper/whisper-diarization, real-time NDJSON streaming, SQLite persistence, TTY audio transport, conformance harness. 107K lines, 2000+ tests, zero unsafe code.

agent-first asr bayesian cli diarization
10
2
100% credibility
Found Feb 27, 2026 at 10 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Rust
AI Summary

A command-line tool that orchestrates multiple speech-to-text engines into a unified interface for transcribing audio files or live microphone input with streaming output, persistence, and adaptive selection.

How It Works

1
🔍 Discover easy audio transcription

You hear about a simple tool that turns spoken words in audio files or recordings into readable text, perfect for meetings or podcasts.

2
📥 Get the tool ready

Download and prepare the tool on your computer so it's all set to use.

3
🎤 Choose your audio

Pick an audio file from your device, pipe in sound, or record live from your microphone.

4
⚡ Start transcribing

Hit go and watch as it smartly picks the best way to convert speech to text quickly and accurately.

5
Pick output style
📄
Quick text

Get a clean written version of what was said right away.

📊
Full details

See timestamps, speakers, and history for deeper insights.

6
💾 Save your results

Everything is automatically stored so you can review past transcriptions anytime.

✅ Perfect transcripts ready

Enjoy accurate, organized text from your audio, ready to read, share, or use in your projects.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 10 to 10 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is franken_whisper?

Franken_whisper is an agent-first Rust ASR orchestration stack that unifies whisper.cpp, insanely-fast-whisper, and whisper-diarization behind a single binary. It solves fragmented speech-to-text pipelines by using Bayesian backend routing across engines, delivering real-time NDJSON streaming for agents, SQLite persistence for run history, and TTY audio transport for low-bandwidth links. Developers get a CLI with commands like `transcribe`, `robot run`, and `tty-audio encode` for instant structured output.

Why is it gaining traction?

It prioritizes agent-first organization on GitHub with machine-readable NDJSON events over human-friendly text, plus adaptive Bayesian routing that picks the best backend per audio clip. The 107k lines of zero-unsafe code, backed by 2000+ tests and a conformance harness, ensure determinism and reliability missing in script-based wrappers. TTY audio with retransmission handles lossy streams, making it ideal for embedded or remote agent workflows.

Who should use this?

Rust developers building AI agents that need reliable, streaming ASR without parsing inconsistent CLIs. Teams orchestrating transcription in real-time apps, like voice agents or meeting analyzers, where persistence and error codes matter. Anyone tired of gluing Python backends manually and wanting a safe, observable alternative.

Verdict

Worth evaluating for agent-first ASR needs—107k lines with 2000+ tests signal serious engineering, despite 10 stars and 1.0% credibility score. Maturity shows in docs and CLI, but expect path dependencies; fork or watch for crates.io release.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.