AI Agent Allowlist
Home Page-Types Database Agent Guardrails 2026 Incidents API Docs Pricing
Resources
Use Cases (15) Industries & Buyers (12) Learn: Core Concepts (12) Implementation Guides (15) Comparisons (8) Schema & Data Reference (6) FAQ Glossary
Why It Matters
2026 Agent Incidents Category Targeting Database Refreshes Contact Customer Login
Download Free Sample
AI Agent Allow List: Cursor

Does Cursor train on your data?

The AI agent allow list checked Cursor against its own published privacy policy and terms of service, 17 September 2026.

Cursor's fields, checked

Every value below comes from the vendor's own published privacy policy, terms of service or terms of use, read and dated. Click show clause to see the exact sentence and its source link.

VendorCategoryTrains on dataOpt-outEnterpriseAPIChecked
Cursor
cursor.com
Agents & AutomationNo
Yes
Yes
Not stated17 Sep 2026

These 7 fields were checked but their values are not shown here at all, only their names. Values are in the full register.

Opt-out covers past submissions: locked Feedback re-opens training: locked Abuse monitoring retention: locked De-identified data still used: locked Staff may read content: locked Zero data retention option: locked Third-party providers may train: locked

What Cursor's own terms say

Cursor's privacy policy states: we do not use Inputs or Suggestions to train our models, or permit third parties to use them for training, unless flagged for security review, explicitly reported as feedback, or explicitly agreed to.

Cursor's terms of service add a stronger enterprise-facing statement: Anysphere will not use content to train, or allow any third party to train, any AI models, unless the customer has explicitly agreed.

What Cursor as an agent sees

Running as a coding agent, Cursor works with repository-level access, which can include configuration files and credentials left in code, beyond the single file a developer has open.

The opt-out for training preferences is managed from inside the Cursor application itself, per Cursor's own documentation.

FAQ

Does Cursor train on our code?

Cursor's privacy policy states it does not use inputs or suggestions to train its models, or permit third parties to train on them, except when content is flagged for security review, explicitly reported as feedback, or explicitly agreed to for training.

Can we opt out of the exceptions?

Cursor's policy states that instructions for managing training preferences for inputs and suggestions are available inside the service itself.

Does the enterprise tier change this?

Cursor's terms of service state that Anysphere, the company behind Cursor, will not use content to train or allow any third party to train any AI models unless the customer has explicitly agreed to that use.

What does Cursor as an agent actually see?

Cursor operates with repository-level access when running as a coding agent, which can include configuration files and any credentials left in code, not only the file open in the editor.

What is still locked on this page?

Seven fields, including opt-out scope, feedback exceptions and monitoring retention, are shown by name only here and unlocked in the full register.

How current is this page?

Every field carries the date it was checked against Cursor's published terms, shown above.

See every Cursor field, unlocked