How to get started with My Architect, what the project model is made of, how to connect an AI agent over MCP — and a set of real use cases showing how teams put it to work. Short and to the point — no filler.
How to sign in, create your first project, and what gets scaffolded when you start a project in My Architect.
What the project model in My Architect is made of — hierarchy, requirements, releases and diagrams.
How to issue a key and connect Claude or another MCP agent to a project's model in My Architect.
Connect Claude Code or Cursor to My Architect's MCP. The agent pulls the next task, claims it, closes it — and your board updates itself.
Agents mark stories done while a blocker or open question is unresolved. My Architect enforces closure at the domain layer so your status stays honest.
Put next quarter's scope change through a pull request — because in My Architect the plan is just git-friendly files, and the file is the source of truth.
Your WBS, story map, and requirements are one tree, not three. Move a card in the map and the breakdown, status, and rollup update — nothing to reconcile.
Turn an Event Storming or C4 board into a gated part of the plan: a DSL an agent authors over MCP, a grammar check, and hotspots that block "done".
USE-CASE doc: give your agent persistent repo memory via the code-graph + recursive-context skills, with honest measured A/B economics.