How it works

A deeper look at the pipeline.

1) Input

Upload an audio/video file or paste a URL. Large files are handled server-side.

2) Transcription

We transcribe (chunking if needed) to produce time-coded segments.

3) Ad detection

An LLM flags spans with clear sponsorship cues (e.g., “sponsored by”, promo codes, or explicit buy/download CTAs).

4) Summary

We generate concise bullets plus key takeaways.

5) Export clean audio

FFmpeg removes ad spans and exports the ad-free result. You can play it and download it.

Start Cleaning Now Back to Home