OpenAI text watermark detector

Approved researchers and expert organizations can use the detector to evaluate text watermarking and its limitations.

Availability

Applications open starting today; initial access is case-by-case for approved researchers and expert organizations.

Pricing model

Not established

Deployment

Not established

Privacy

The tool reports whether it detects an OpenAI watermark without identifying users or revealing prompts or conversations.

Licence

Not established

• Approved researchers and expert organizations can apply for detector access starting today. ^[1]

  * **Providing detector access to researchers and expert organizations.** Approved researchers and expert organizations can [apply](https://openai.com/form/content-provenance-api/) starting today.

• Detector access will initially be granted case by case. ^[1]

  In accordance with the [Code of Practice⁠(opens in a new window)](https://digital-strategy.ec.europa.eu/en/policies/code-practice-ai-generated-content), access will be initially granted on a case-by-case basis to support evaluation and improvement of text provenance.

• The detector reports whether it detects an OpenAI watermark. ^[1]

  Our detector looks for that signal to assess whether a passage contains an OpenAI watermark.

• The detector does not identify the user or reveal prompts or conversations. ^[1]

  * **A watermark does not identify the user.** It does not associate a person, organization, account, prompt, or conversation with the text.

• OpenAI says the detector is not publicly available at launch because of risks from missed watermarks and false positives. ^[1]

  Given the risk of missed watermarks and false positives, we are not making it publicly available at launch.

• Build and govern workplace agents with North 2 Daily brief 2026-10-05