Last updated: July 30, 2026

AI Service Terms

These Terms apply to your use of Craepas’s AI conversations, task dispatch, skill execution, file analysis, browser collaboration, artifact generation, and automation capabilities.

1. Nature of AI Outputs

AI outputs are drafts or auxiliary results generated based on user input, workspace context, and model capabilities, and may contain errors, omissions, or outdated or incomplete content, or be unsuitable for a particular purpose. You should exercise your own judgment and perform human review before use.

2. High-Impact Scenarios

For legal, tax, financial, medical, hiring, credit, insurance, educational assessment, public safety, contractual commitment, regulatory filing, or other high-impact matters, AI outputs must not be relied upon as the sole basis. You must have them reviewed by qualified or authorized personnel.

3. Responsibility for User Input

You must ensure that the materials you input, upload, or authorize for processing are lawful and accurate, and that you have obtained the necessary rights or consents. Do not submit unlawful content, malicious code, unauthorized personal information, trade secrets, or specially regulated data.

4. Boundaries of Automated Execution

When you authorize AI to perform browser, file, connector, or third-party system operations, you should confirm the scope and consequences of those operations. Actions involving external publishing, payments, deletion, submission, signing, or other acts that cannot easily be undone should be confirmed by a human first.

5. Content Rights

To the extent permitted by law and without infringing third-party rights, you may use the output content generated by the services that belongs to you. You remain responsible for the consequences of using the output content.

6. Model and Service Limitations

Model capabilities may change with versions, providers, network conditions, quotas, or security policies. We may impose limits on request frequency, context length, file size, task duration, connector permissions, and high-risk operations.