Categories › AI & Agents › Multi-agent, orchestration & workflows
Free agent tools, no sign-in: browser flows, test inboxes, web extract, site and security scans.
Remote server: your requests and data go to gluo.dev. Claude Code: claude mcp add --transport http gluo https://gluo.dev/mcp · Claude app: Settings → Connectors → Add custom connector.
The tools this server listed when we connected to it, in its own words.
create_job — Hand off a multi-step, multi-day task that golive runs to completion, even after this session ends. Paid or irreversible steps pause for theget_job — Current status of a job, what is blocking it, and what to do next.list_jobs — All jobs, newest first.undo_job — Reverses finished steps that can be undone, newest first. Purchases that can't be refunded are listed, not undone.list_services — Services Gluo runs for agents (site checks, screenshots, domain availability, email setup, dependency vulnerabilities, ...), each with its erequest_budget — Request a spending budget for specific operations. Ask only for the operations and amount the task needs: the user sees the exact scope, capget_budget — Status, spend, remaining amount and per-operation usage of a budget.call_api — Make one call. Leave out budgetId to use the free tier (free operations only, daily allowance). Input must match the operation's schema exacclose_budget — End a budget when the task is done. Returns a signed statement of every call and charge.request_service — If a task needs a service, API or action Gluo doesn't offer, describe it in a sentence or two. This is how Gluo decides what to build next. Not tested by KitScout yet. When our harness runs it, pros, cons and the tasks we tried appear here, with the date and evidence.