GitHub radar
Automatically find out who spoke and when in a recording
This model listens to audio with multiple people and figures out who spoke and exactly when — no manual labeling needed.
It's a program that splits a conversation recording by voice: it shows who spoke and when, but doesn't transcribe the actual words.
When it helps
Useful if you edit podcasts, videos, or calls with several participants. If you need the actual text, this alone isn't enough — you'll need a separate speech-recognition step.
Pros
- Tells apart up to 8 speakers in one recording
- Runs on an ordinary laptop, no GPU needed
- Fastest mode reacts in 0.32 seconds
- Full-recording error rate is a third lower than before
Cons
- Only recognizes voices, not actual words
- Getting a text transcript needs an extra step
How to set it up — step by step
- 1Open your AI agent (Claude Code or Codex) in your project folder.
- 2Send the agent the request text shown below.
- 3Prepare a test recording with several speakers in WAV format.
- 4Check that the agent installed the model and NeMo-Speech.cpp without errors.
- 5Review the result — a list of segments like "speaker 1: start–end".
- 6If you need the transcript text, ask the agent to add speech recognition as a separate step.
Text for your agent
Copy this and send it to your agent — Claude Code, Codex, any of them:
Set up the local speaker diarization model Nemotron 3 Diarization from NVIDIA: https://huggingface.co/nvidia/Nemotron-3-Diarization - install it via NeMo-Speech.cpp, run it on a sample multi-speaker recording and show me who spoke when.
A regular laptop with 8 GB of RAM or more - the model is about 100 MB and needs no GPU.
Open on Hugging Face▌ More finds