Hi everyone It seemed some developers dont know how fairy reviews work (https://code.ffmpeg.org/michaelni/Fairies)
Fairy is written in python, its currently 3 processes 1. one agent per repository (fetching from repo, caching, filling json based file db) 2. workers which pick up queued tickets from the file db and perform the review using multiple LLMs and multiple (remote) containerized test environments. 3. fairy-ui, the part that monitors this and allows a human to approve / redo / skip reviews (or just auto approve) Heres a screenshoot from current fairy-ui https://ffmpeg.org/~michael/fairy-ui-2026-08-23.png the UI part is merly a viewer on top of the file db, but one could run it without a LLM just as a pure viewer of PRs/Issues (which should work without a internet connection once the tickets are fetched) thx -- Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB z(9) = an object that transcends all computable functions describable in finite terms. - ChatGPT in 2024
signature.asc
Description: PGP signature
_______________________________________________ ffmpeg-devel mailing list -- [email protected] To unsubscribe send an email to [email protected]
