One backend for your apps' ops.
Point your web and desktop apps at cmdline. Read it from the CLI, or let your coding agent run it over MCP — one API key for all of it.
What it runs
Uptime, errors, and deploys share one timeline; desktop distribution and agent control sit on the same key.
Uptime monitoring →
HTTP/HTTPS checks from outside your infra. Intervals from 30 seconds to an hour. Response-time tracking, uptime percentages, and alerts to ntfy / Slack / email / webhook.
Error tracking →
Capture exceptions from your app. Auto-grouping by fingerprint, deduplication, stack traces. Spike detection on Pro+.
Deploy tracking →
Record deploys with commit SHA and status. Errors and incidents show on the same timeline, so you see cause and effect against what shipped.
Desktop distribution →
Host Sparkle appcasts, ship releases, and query your install base and update adoption. Dogfooded on TmpDisk — 616 live installs.
Operated by agent →
An MCP your agent uses to provision monitors, mint scoped keys, wire alerts, cut releases, and account for incidents.
Incident accountability →
When your agent ships a regression, it files who, why, and the fix against the incident — a lean postmortem, on the record.
Your agent operates it.
cmdline exposes an MCP — a standard interface your coding agent (Claude Code, Cursor, and others) speaks. The read side answers questions: is anything down, what errors are new, what shipped, how is adoption. The write side lets the agent set things up: register an app and mint its key, wire a monitor, route alerts to a channel.
Write access is scoped per key and can't escalate, so you can hand an agent exactly the permissions it needs — and no more.
Built for the terminal, too
cargo install cmdline-cli, then one command for a read on your whole business — or leave cmdline top on a screen. Ship deploy markers from CI, tail errors, pull status from whatever workflow you already have.
Ready when you are.
Free for one app. Mint a key, add the MCP, and ask your agent if your business is healthy.