Everything you need to hire your first worker and keep it on the record.
You never have to open the dashboard. Install WorkerOS where your team already talks and hire, approve, and redirect from there.
WorkerOS speaks MCP. Add it to Claude Code, Cursor, Codex, or any client and drive workers from where you already work.
$ npx -y @floomhq/workeros mcp install --target claudeShips the `workeros-mcp` stdio server from the `@floomhq/workeros` package.
Then: "run client-follow-up for the Acme call" from your agent. The run lands on the record like any other.
Drop in SOPs, pricing sheets, tone guides, links. Every worker brief comes pre-loaded with what your team already knows, and each run logs which files it used so you can audit the why behind every output.
Anything that leaves the building (emails, CRM writes, posts) is held for your yes by default. Approve in Slack, WhatsApp, or the app. You can relax the gate per worker once you trust it.
Any. Bring your own keys or use the defaults. Model choice is per worker.
Encrypted at rest, used only for the workers you build. Revoke at the source any time.
Yes. Workspaces share workers, brain, and the run record.
WorkerOS Cloud is in early access. Pricing is being finalised with the first teams.