“model providers (including Cursor) may run risk classifiers to detect violations of terms and usage policies, and if your prompts or conversations trigger abuse detectors your data may be stored for investigation...”
This analysis describes what Cursor's agreement states, permits, or reserves. It does not constitute a legal determination about enforceability. Regulatory applicability and practical outcomes may vary by jurisdiction, enforcement context, and individual circumstances. Read our methodology
This establishes an explicit exception to ZDR commitments: abuse-triggered data may be retained by model providers or Cursor outside the standard zero-retention framework.
The updated terms clarify that Cursor maintains zero data retention agreements with all AI model providers and will not use customer data for training. However, the revised language now explicitly states that model providers may run risk classifiers to detect policy violations, and if your prompts or conversations trigger abuse detectors, your data may be stored temporarily for investigation and deleted according to the providers' retention policies. The policy removed the legacy Privacy Mode option; non-ZDR models now require explicit admin opt-in. You can review the specific retention policies of OpenAI and Anthropic through their documentation.
View change record →If the reader's prompts or conversations trigger an abuse detector, that data may be stored for investigation by model providers or Cursor, notwithstanding any ZDR agreements.
How other platforms handle this
We may retain some of your Data even after you cease to use our Services, for instance if the Data is necessary to meet our legal obligations, such as tax and accounting requirements.
Where specific retention periods are required by law, we comply with those timeframes; otherwise, we apply the criteria above to determine how long to retain the information.
If the Services are made available to you through an organization (e.g., your employer), we retain your information as long as required by the administrator of your account.
"model providers (including Cursor) may run risk classifiers to detect violations of terms and usage policies, and if your prompts or conversations trigger abuse detectors your data may be stored for investigation...Excerpt from Cursor's Data Use & Privacy Overview
Get the weekly coverage letter
Companies change their terms without announcing it. We capture governance documents, compare each new version with the last, and publish what changed, with the source text behind it. One email a week on what we found and what we could not reach.
This establishes an explicit exception to ZDR commitments: abuse-triggered data may be retained by model providers or Cursor outside the standard zero-retention framework.
If the reader's prompts or conversations trigger an abuse detector, that data may be stored for investigation by model providers or Cursor, notwithstanding any ZDR agreements.
ConductAtlas has identified this type of provision across 268 platforms. See the full comparison.
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Cursor.