Nothing of ours
can go down.
Codescry scans, scores and stores everything on your machine — no accounts, no API, no server of ours in the path. So this is not an uptime page for a service. It is a live check of the few third-party endpoints the app touches, run from your browser, right now.
Read these results as "reachable from here", not "the service is up". The checks run in your browser, so a corporate proxy, a captive portal or a content blocker produces exactly the same red bar as a real outage. History is stored in this browser only (localStorage, last 40 checks) — nothing is reported back to us, because there is nowhere to report it to.
The parts that actually decide whether a scan runs
These live inside your machine, so no web page can measure them — the app checks them itself and blocks a run with a plain reason when one is missing (Spec §5.2, model gate).
Every analysis call goes here. Not running means the run never starts, with a fix-it prompt instead of a stack trace.
curl http://127.0.0.1:11434/api/tags
Onboarding pulls the model and verifies it with a test prompt. After that the model host can vanish and scanning is unaffected.
ollama list
Findings, scores and history. Settings shows the exact path, exports everything to JSON, and deletes it on request.
Settings → Data → Open data folder
What can go wrong is a bad build, not an outage: a release that fails to install, a broken model pull, a scan that hangs. Those are tracked in the open on GitHub, and a fix ships as a new version rather than a status update.