AI Policy

Hound is not vibe-coded. I started developing Hound years ago and even made a reddit post about it. Since then, a lot has changed in the software development landscape.

Here's my current development approach:

• I use AI to throw around ideas, and have a discussion on potential solutions to implement a new feature, or architectural decision
• I often disagree with AI's opinions. eg. Since I know the project best, there are times to make practical decisions over theoretically 'correct' ones
• Once I've decided on a plan, I code everything myself.

Actual coding is rarely the development bottleneck in Hound, when you plan well, the actual code changes should be short and surgical. I know the codebase better, and I've found that using AI to code doesn't speed up the process. With AI, the time spent debugging offsets any time saved.

I don't use AI to write for me either, everything in the documentation, repos, are my own writing. Pinky promise.

Contributing

Right now, I'm the only one working on Hound, so contributors are very welcome!

Feel free to report bugs on Github, or make a pull request.

General Guidelines

• Don't vibe-code, you need to understand every line of code in a PR. I recommend following my approach in the AI section above.
• PRs should be as surgical as possible. Fix or add one thing at a time.
• If you'd like to add a new feature, especially a major one, please discuss this with me first, or open an issue.
• You're free to make third-party clients and tools for Hound, refer to the API Documentation in this site. We encourage these to be open-source as well!
• Feel free to contact me for any help, such as understanding the codebase, help setting up the dev environment, etc.