How ReelAudit reverse video search works

No major public search engine accepts a full video file as a query. Every serious reverse video workflow — including ours — extracts distinctive frames (and sometimes audio) and searches those against image and media indexes.

1. Keyframe extraction

On this MVP, frames are sampled in your browser with the HTML5 video API. We take several frames across the timeline (up to five on Free). Later versions can add scene-change detection via ffmpeg on the server.

2. Multi-engine matching

Google Lens, Bing Visual Search, Yandex Images, and TinEye cover different corners of the web. Overlap is imperfect — that is why cross-checking matters. TinEye's oldest-first sort is especially useful for copyright timelines.

3. Timeline & evidence

Results are grouped as candidate sources, not isolated frames. A "Likely original" badge means earliest public date among returned candidates — not a courtroom verdict. Exportable evidence packs and human review sit on top of this board.

What we will not claim

• A match is always the original author.
• No match means the clip is exclusive or authentic.
• Automated DMCA filing without your review.
• Face search or biometric identification.

Ready to try? Open the search tool or request a human report.