Interesting project. Monitoring how ChatGPT, Gemini, and Claude cite a site over time feels genuinely useful, especially since most teams currently have to rely on scattered manual checks. I also like that Canonry is self-hosted and SQLite-based instead of being another black-box dashboard. Curious how you handle query selection and repeatability so the visibility scores stay comparable over time.
By query selection do you mean what you track? That is configurable by the user (adding support for auto generation soon).