README.md (1534B)
1 2 # notecrumbs 3 4 [](https://github.com/damus-io/notecrumbs/actions) 5 6 A nostr opengraph server build on [nostrdb][nostrdb], [egui][egui], and 7 [skia][egui-skia]. It renders notes using the CPU in around 50ms. 8 9 [nostrdb]: https://github.com/damus-io/nostrdb 10 [egui]: https://github.com/emilk/egui 11 [egui-skia]: https://github.com/lucasmerlin/egui_skia 12 13 14 ## Status 15 16 WIP! 17 18 - [x] Local note fetching with nostrdb 19 - [x] Basic note rendering 20 - [x] Fetch notes from relays 21 - [ ] Render profile pictures 22 - [ ] Cache profile pictures 23 - [ ] HTML note page 24 25 Very alpha. The design is still a bit rough, but getting there: 26 27 <img style="width: 600px; height: 300px" src="https://damus.io/nevent1qqstj0wgdgplzypp5fjlg5vdr9mcex5me7elhcvh2trk0836y69q9cgsn6gzr.png"> 28 29 ## Relay discovery & metrics 30 31 - Notecrumbs keeps long-lived relay connections and now learns new relays from every event it ingests. Relay list (`kind:10002`) events and contact lists (`kind:3`) are parsed for `r`/`relays` tags as well as per-contact relay hints, and the pool deduplicates and connects to any valid URLs it sees. 32 - Per-request fetch loops feed those hints straight into the shared pool, so visiting a profile helps warm future requests that need the same relays. 33 - Relay pool health counters (ensure calls, added relays, connect successes/failures, and active relay count) are exposed via Prometheus at `http://127.0.0.1:3000/metrics`, and a mirrored summary is logged every 60 seconds.