← Back to all providers

Cursor

Cursor has notably strong AI training protections - they explicitly won't use your code to train AI models unless you agree. However, they require binding arbitration for disputes, disclaim most warranties for their AI suggestions, and can terminate your account for any reason. The auto-code execution feature puts all risk on you if the AI-generated code causes problems.

TOS reviewed:May 2026
View Original TOS

The Bottom Line

Cursor has notably strong AI training protections - they explicitly won't use your code to train AI models unless you agree. However, they require binding arbitration for disputes, disclaim most warranties for their AI suggestions, and can terminate your account for any reason. The auto-code execution feature puts all risk on you if the AI-generated code causes problems.

Get notified when Cursor updates their TOS

We'll email you what changed and what it means.

By subscribing, you agree to our Terms and Privacy Policy.

Notable Clauses

Auto-Code Execution Risk1.7. Auto-Code ExecutionVERIFIED

YOU ARE SOLELY RESPONSIBLE FOR ANY IMPACT RESULTING FROM USE OF THIS FEATURE, INCLUDING ENSURING APPROPRIATE SAFEGUARDS, TESTING, AND MONITORING ARE IN PLACE.

Translation: If their auto-running AI code breaks your system or deletes data, that's entirely on you.
VERIFIEDView proof →
Rate it (1 amazing → 10 crazy):
Anonymous · Privacy
Forced Arbitration16.1. Agreement to ArbitrateVERIFIED

You agree that any and all disputes or claims that have arisen or may arise between you and Anysphere, whether arising out of or relating to these Terms (including any alleged breach thereof), the Service, any advertising, or any aspect of the relationship or transactions between us, will be resolve...

Translation: You can't sue them in court - all disputes must go through private arbitration.
VERIFIEDView proof →
Rate it (1 amazing → 10 crazy):
Anonymous · Privacy
Class Action Ban16.2. Prohibition of Class and Representative ActionsVERIFIED

YOU AND ANYSPHERE AGREE THAT EACH OF US MAY BRING CLAIMS AGAINST THE OTHER ONLY ON AN INDIVIDUAL BASIS AND NOT AS A PLAINTIFF OR CLASS MEMBER IN ANY PURPORTED CLASS OR REPRESENTATIVE ACTION OR PROCEEDING. UNLESS BOTH YOU AND ANYSPHERE AGREE OTHERWISE, THE ARBITRATOR MAY NOT CONSOLIDATE OR JOIN MORE ...

Translation: You can't join with other users to sue them as a group - only individual claims allowed.
VERIFIEDView proof →
Rate it (1 amazing → 10 crazy):
Anonymous · Privacy
Termination for Any Reason9. TerminationVERIFIED

If you have a Subscription Service, we may terminate the Subscription Service at any time for any other reason.

Translation: They can cancel your paid subscription whenever they want for any reason.
VERIFIEDView proof →
Rate it (1 amazing → 10 crazy):
Anonymous · Privacy
No Warranties on AI Suggestions14. DISCLAIMER OF WARRANTIESVERIFIED

THE SERVICE AND SUGGESTIONS ARE PROVIDED "AS IS" AND ON AN "AS AVAILABLE" BASIS. ANYSPHERE DISCLAIMS ALL WARRANTIES OF ANY KIND, WHETHER EXPRESS OR IMPLIED, RELATING TO THE SERVICE AND SUGGESTIONS, INCLUDING: (a) ANY IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, TITLE, QUIET...

Translation: They won't guarantee their AI suggestions work or are safe - use at your own risk.
VERIFIEDView proof →
Rate it (1 amazing → 10 crazy):
Anonymous · Privacy
Beta Services Liability Waiver1.6. Beta ServicesVERIFIED

ANYSPHERE SHALL HAVE NO LIABILITY WHATSOEVER ARISING OUT OF OR IN CONNECTION WITH BETA SERVICES - USE AT YOUR OWN RISK.

Translation: For any beta features, they have zero responsibility if something goes wrong.
VERIFIEDView proof →
Rate it (1 amazing → 10 crazy):
Anonymous · Privacy
You Own Your Content5.3. ContentVERIFIED

You retain all of your right, title, and interest that you have in Inputs, and Anysphere hereby assigns to you all of our right, title, and interest if any in and to any Suggestions.

Translation: You keep ownership of your code and they even give you rights to the AI suggestions.
VERIFIEDView proof →
Rate it (1 amazing → 10 crazy):
Anonymous · Privacy
No AI Training Without Consent1.3. Model TrainingVERIFIED

ANYSPHERE WILL NOT USE CONTENT TO TRAIN, OR ALLOW ANY THIRD PARTY TO TRAIN, ANY AI MODELS, UNLESS YOU'VE EXPLICITLY AGREED TO THE USE OF CONTENT FOR TRAINING.

Translation: They actually won't train AI on your code unless you specifically opt in - rare and good.
VERIFIEDView proof →
Rate it (1 amazing → 10 crazy):
Anonymous · Privacy

Protect Yourself

ACTIONABLE

Settings and actions you can take to limit your exposure

HIGH IMPACT

Disable AI Training

Turn off the option that lets them use your code inputs and AI suggestions for training models

How to do it:
  1. Open Cursor settings
  2. Look for AI training or data usage preferences
  3. Disable any options related to using your content for model training
HIGH IMPACT

Disable Auto-Code Execution

Turn off the feature that automatically runs AI-generated code without your review

How to do it:
  1. In Cursor settings, find the auto-code execution feature
  2. Disable automatic execution of AI suggestions
  3. Ensure you manually review all code before running
MEDIUM IMPACT

Backup Your Code Regularly

Since they can terminate service at any time, keep local copies of your work

How to do it:
  1. Use git to maintain local repositories
  2. Export or download your projects regularly
  3. Don't rely solely on Cursor's cloud storage
MEDIUM IMPACT

Opt Out of Arbitration

You have 30 days to reject their forced arbitration clause and preserve your right to sue

How to do it:
  1. Email legal@cursor.com within 30 days of creating your account
  2. Include your full name and clear statement that you opt out of mandatory arbitration
  3. Send from the email address used for your account
LOW IMPACT

Monitor Subscription Billing

They auto-renew subscriptions and require 24-hour notice to cancel

How to do it:
  1. Set a calendar reminder 2-3 days before renewal
  2. Cancel through billing menu or email hi@cursor.com
  3. Verify cancellation was processed before renewal date

Better Alternatives in AI

These ai services scored lower (less crazy) on their terms

Disclaimer:The clauses shown above are excerpts from publicly available Terms of Service documents. Ratings and severity levels reflect community and AI assessment, not legal advice, factual claims, or determinations about a company's legality or intent. Clauses marked “VERIFIED” have been confirmed as exact quotes from the source document. Clauses marked “AI-extracted” were identified by AI and may contain errors. Always verify against the original TOS.

Get notified when Cursor updates their TOS

We'll email you what changed and what it means

By subscribing, you agree to our Terms and Privacy Policy.