← All stories
● Covered by 2 sources · 2 reportsMedium impact1 negative1 neutral

OpenRouter Removes Providers and Endpoints Due to Data Policy Violations and Prompt Theft

🔄 Updated 1h ago
New to BrevFeed? We gather this story from every outlet covering it into one summary — ranked by real-world impact, not just the latest headline — so you never miss what matters. What is BrevFeed? →

Key points

  • OpenRouter removed three inference providers from its platform.
  • 125 endpoints lost Zero Data Retention (ZDR) eligibility.
  • Security researchers reported data policy violations and prompt theft.
  • Incidents involved attackers stealing data from providers since mid-July.
  • OpenRouter itself was not breached; affected customers were notified.

Provider Off-boarding and Endpoint Removal

OpenRouter has removed three inference providers from its platform following reports from security researchers regarding data policy violations. Additionally, 125 endpoints have been stripped of their Zero Data Retention (ZDR) eligibility. These actions were taken after investigations confirmed issues with how certain providers handled data.

Data Policy Violations and Prompt Theft

Security researchers reported that some inference providers had data compromised or failed to uphold their data policies. The findings indicated that attackers stole data from within three AI inference providers for several months, with the earliest breach starting in mid-July. This included prompts that were marked for Zero Data Retention, which are intended to never be stored.

No Breach of OpenRouter

OpenRouter stated that the findings did not show any breach or unauthorized access of its own systems. The incidents were isolated to the inference provider level. OpenRouter has notified all potentially affected customers about these data processing incidents.

Canary Token Research Method

The discovery of the prompt theft came from canary token research conducted by an industrial cybersecurity company. Between July and November, approximately two million prompts were seeded with unique web links and credentials. Any use of these planted secrets indicated that the prompt carrying it had been read, revealing the data compromise.

Ongoing Measures and Future Validation

OpenRouter is adapting its testing methods and will require validation from providers before restoring ZDR eligibility. This ongoing work aims to ensure that data policies are consistently upheld across its platform and to prevent future incidents of this nature.

✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →

The daily brief

One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.

One email a day. Unsubscribe in one click, any time.

Today's brief

Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.

~4 min · 3 stories · Nov 21

▶ Play today's brief Listen on Spotify

New every morning, and the back catalogue is archived by date.

How outlets covered it

OpenRouter has removed three AI inference providers from its platform and notified affected customers after researchers discovered months of prompt theft. Attackers stole customer data, including prompts marked for Zero Data Retention, from these providers without their knowledge, indicating a breach at the inference provider level rather than OpenRouter itself.

OpenRouter off-boarded three inference providers and removed 125 endpoints from Zero Data Retention (ZDR) eligibility after security researchers reported data policy violations by some providers. The company is implementing ongoing testing and requiring validation before restoring ZDR eligibility to ensure data policies are upheld across its platform.