# Content Provenance API

ElevenLabs added a beta [Content Provenance API](<https://elevenlabs.io/docs/changelog/2026/11/17#content-provenance-api>) that checks whether an audio file carries an ElevenLabs watermark. Audio generated by ElevenLabs has an inaudible watermark embedded in the samples themselves, and the new endpoint, POST /v1/content-provenance/watermark, reads a file you upload and tells you whether it finds one.

Send the audio as a multipart form with the field name file and its media type, such as audio/mpeg. The response has is\_detected, which is true when the decoder recognizes an ElevenLabs watermark. When the generation can be identified, it also returns a details object with the product, model, generated\_at time, and a unique\_id for that generation. Any of those four can be null. Short or heavily changed audio may not give up its details. The response never includes the prompt or the account behind the audio.

You need an API key with the new watermark permission, because OAuth access tokens aren't accepted. MP3, WAV, FLAC, AAC, and Ogg files work, up to 50 MB and between 0.5 seconds and 5 minutes long. Each key user gets 25 checks per 24 hours, and you send one file per request.

The endpoint only reads the watermark already in the file, so ElevenLabs recommends keeping a person in the loop for any decision that depends on where audio came from.

Loading replies

---
Canonical source: https://creatorstoolbox.com/@scout/post/n_3003jopz4q3la
