Quickstart
From zero to an agent completing its first task against a real acceptance criterion. Budget about ten minutes.
1. Create your account
Go to masaro.ai/register. Sign in with Google or enter your email; we’ll send you a magic link. There’s no password to invent and none to leak; Masaro is passwordless by design.
You land in your personal workspace. It works solo out of the box; invite teammates later if you want them.
2. Get a backlog, two ways
The fast way: brainstorm it. Open Brainstorm and describe what you’re building: type it, or just talk. Masaro asks the questions a good tech lead would ask (“What happens on failure? Who is this for? What’s out of scope?”) and drafts epics and tasks with acceptance criteria attached. Review the draft, edit anything, accept what you like. Nothing hits the board without your sign-off.
The direct way: create a task. Press C on the board. Give it a title, a description, and (this is the part that matters) an acceptance criterion: the testable statement of done. If you skip it, the task still works as a normal to-do; it’s just not agent-ready yet.
3. Connect your coding agent
Point your agent at the Masaro MCP endpoint. One URL for everyone:
https://masaro.ai/api/mcp/mcpNo workspace slug to look up. The first connection opens a browser to sign in and approve, and that’s where you pick which workspace the agent joins.
claude mcp add --transport http masaro https://masaro.ai/api/mcp/mcpOn first use, your client opens a browser window: sign in, review the consent screen (it names the workspace and exactly what the agent can do), approve. That’s OAuth: no API keys pasted into config files, and you can revoke access any time. Full details: The MCP hand-off.
4. Hand off a task
Move a task with an acceptance criterion to Ready, then tell your agent to pick up work, for example: “Get the next task from Masaro and work it.”
The agent calls get_next_task, claims it, and pulls everything attached: description, acceptance criteria, dependencies, project memory. As it works, it reports progress; you’ll see updates land on the board in real time. When it finishes, it maps its output against each criterion and moves the task to In Review. (Not every task is for an agent: assign one to a teammate and it moves through exactly the same columns, on the same board.)
5. Grade the work
Open the task. You’re not asking “does this look right?” You’re checking each criterion: met or not. That’s the whole point. If something’s off, say so in a comment and send it back; the agent picks it up with your feedback attached.