Available on macOS, Windows, and Linux · Free for individuals
Move beyond one prompt at a time. Optamus helps you delegate development tasks to AI agents, work across your development environment, and review their changes before accepting them.
Sample workspace shown — install Optamus to run it on your own code.
Start it where you are. The desktop app, the CLI, and the API are the same agent with different verbs.
Parallel runs sit side by side, each with its own context. The diff, the log, and the interrupt button are always one keystroke away.
Agent thread
workspaces: 4 shards — unit, integration, auth-real, contractOne command starts a run. Flags control parallelism, output format, and where the log lands — so it works in a script or a pipeline.
$ optamus run "migrate to vitest" --parallel 4 --json ▸ 4 runs queued · worktrees ready ✓ run/1a2b 62 tests +214 −188 ✓ run/1a2c 62 tests +96 −96 ✓ run/1a2d 62 tests +18 −0 ✓ run/1a2e 62 tests +0 −214
Start a run with a POST, subscribe to its events over server-sent events, and collect a single receipt when it settles.
$ curl -N https://api.optamus.cloud/v1/runs \ -H "Authorization: Bearer $OPTAMUS_KEY" \ -d '{"task":"triage open issues","limit":20}' event: run.started run/7c4d event: step.completed tests green event: run.settled {"state":"completed"}
This release fans out runs in the CLI and cuts cold start in half.
Optamus 1.4 lets a single command fan out into independent agent runs with separate worktrees. The scheduler, the limits, and what we learned.
Server-sent events, resumable runs, and a settle phase. The decisions behind the Optamus API surface.
Install the desktop app or the CLI, run one command, and review the first diff.