PurifyAI API
Strip C2PA content credentials, AI generator tags and EXIF/IPTC metadata from PNG and JPEG files over HTTP. One endpoint, one header, no SDK.
https://www.purifyai.app/api/v1/scrubcurl -X POST "https://www.purifyai.app/api/v1/scrub" \-H "Authorization: Bearer $PURIFYAI_KEY" \-F "file=@photo.png" \-F "mode=all" \-D headers.txt \-o photo_clean.png# The report is in the response headersgrep -i "X-PurifyAI" headers.txt
What it does
The API rewrites the container of an image, removing metadata segments while leaving the pixels untouched. It is the same engine the browser Studio runs, exposed for pipelines that need it server-side.
- C2PA / Content Credentials — the provenance manifests attached by Adobe, OpenAI and others.
- AI generator tags — the PNG text chunks and XMP blocks that name Midjourney, DALL·E, Gemini and similar.
- EXIF, IPTC and XMP — camera, location and editing history.
What it does not do
The API rewrites metadata only. It does not remove visible watermarks burned into pixels, and it does not process video — both of those are Studio features and run locally in the browser, never on a server.
Only PNG and JPEG can be rewritten server-side. Anything else returns 415 unsupported_format.
Before you start
- API access requires the Agency plan.
- Create a key in your workspace. It is shown once and cannot be retrieved afterwards.
- Requests count against the same monthly quota as the app — 5,000 files per month on Agency.
Base URL
https://www.purifyai.app/api/v1
All requests must use HTTPS. Plain HTTP is redirected, and a redirected POST loses its body in most clients — so point at the canonical host directly.