A plain HTTP API.

Send a file — or a link — and get a verdict with the confidence behind it. Works for images and video; audio is under maintenance and coming soon.

Get your free API key

100 free checks, 30 requests/min — no card required. The key is shown once and stored only as a hash; when the free quota runs out you'll get HTTP 402 and can upgrade from the pricing page. Want to see your keys and usage later, or revoke one? Create an account instead — same free key, plus a page to manage it.

Built for trust, measured for real.

Every number below comes from held-out testing — generators the model never saw in training, photos from sources it never learned from.

92.7%
of AI files caught on the live API bench (112 fresh files, Sep 2026)
0.25%
real photos wrongly flagged (800 photos from unseen sources, Aug 2026)
0.6s
image verdict, server median (Sep 2026)
14
generator families held out in the August 2026 calibration
August 2026 recalibration and September 2026 live-API bench. When the signal is ambiguous, Tayanch says "uncertain" — never a confident guess.
Images, video, and audio
One endpoint family handles stills, clips, links, and audio — each response includes the verdict, a probability, and a confidence. (Audio is under maintenance — coming soon.)
Link analysis
POST a URL to /analyze/url and the server fetches the media (SSRF-guarded) and runs the same pipeline — social previews resolve automatically.
Calibrated thresholds
Verdicts are tuned per use case, so you can trade recall for a lower false-positive rate.
Simple integration
REST over HTTPS, JSON responses, and API-key auth via the X-API-Key header.