The Loop project has strict rules for AI usage.
All AI usage in any form must be disclosed. You must state the tool you used (for example, Claude, ChatGPT, GitHub Copilot, Cursor) and describe how the work was AI-assisted. If we suspect that you have automatically used a tool to open a pull request-such as when the code fails in obvious ways, or when your responses to review comments do not feel human-we may try to verify that you are a human contributor through specific questions. If you fail to respond, refuse to engage, or cannot reasonably demonstrate understanding, your pull request may be closed and further actions may be taken in line with this policy and the broader policies of the Loop repository.
AI-assisted pull requests must be fully verified by a human. You are responsible for ensuring that AI-suggested changes actually work. You must manually review all AI-generated code rather than pasting unreviewed output, and you must not submit code for platforms, environments, or configurations you cannot personally run or test. Pull requests that clearly contain untested or hypothetical code may be closed without detailed review.
Issues and discussions may use AI assistance but must have a full human in the loop. You must have read and understood everything you are posting, edited AI output for accuracy, relevance, and brevity, and verified any claims of fact or provided citations where appropriate. AI is very good at being overly verbose and including noise that distracts from the main point, so you are expected to do the work of trimming this down before submitting.
No AI-generated media is allowed. Only text and code are acceptable AI-generated content, and only when they comply with the rest of this policy.
These rules apply only to outside contributions to Loop. Maintainers are exempt from these rules and may use AI tools at their discretion. They are expected to exercise good judgment, validate AI output, and remain accountable for changes they approve or merge.
The following are not allowed under this policy:
- Undisclosed AI usage in any part of a contribution.
- Pull requests with AI-generated code that is clearly untested, does not build, or targets platforms or environments you cannot run yourself.
- AI-generated media of any kind, including images, diagrams, audio, and video.
- Bulk-dumped AI output with little or no human editing.
- Using AI to repeatedly reframe ideas that have already been rejected by maintainers.
- Automatic or scripted generation of pull requests or comments that you do not personally review, understand, and stand behind.
Before submitting, confirm that:
- You have disclosed any AI usage, including which tools were used and how you used them.
- Your pull request references an accepted issue, unless maintainers have explicitly allowed otherwise for trivial fixes.
- You have manually reviewed and edited all AI-generated content.
- You have run the relevant tests or manual checks appropriate for your changes.
- You have not included any AI-generated media.
- You understand and can explain any code or text you are submitting.
Loop is maintained by humans.
Every discussion, issue, and pull request is read and reviewed by people, sometimes with the help of tools. That review time is limited and valuable. When low-effort or unqualified work is submitted-especially large amounts of AI-generated content-it shifts the burden of validation onto maintainers.
AI can only be useful here when the person using it takes responsibility for the output: understanding what it does, checking that it is correct, and presenting it clearly. If you use AI, you are still fully responsible for the quality, correctness, and tone of everything you submit.
Treat maintainers’ time with respect by doing the hard work yourself: understand the codebase, validate AI suggestions, and send focused, high-quality contributions.
Some parts of Loop have been developed with the help of AI, and maintainers may actively use AI tools as part of their workflow. As a project, we consider AI to be a valuable tool.
This strict AI policy is not anti-AI. It exists because many people currently using AI lack the experience or context to validate its output. The main risk is not the tools themselves, but unqualified usage that produces incorrect, noisy, or misleading contributions.
We include this section to be transparent about how AI is used in Loop, for people who may disagree with it, and to address the misconception that this policy is anti-AI in nature.
This AI policy takes inspiration from the Ghostty project's AI policy and has been tailored specifically for this project. Any adjustments or interpretations are made solely by the maintainers of Loop and do not reflect the positions or policies of the Ghostty project.