Your AI is only as good as what it knows.
The repo holds the code. Scope Architect holds everything else your AI needs: the context it reads before it starts, the memory that survives between sessions, and the docs it reads and writes as it builds. Built into the workspace, not bolted on like a separate tool.
You manage what it knows. It builds from that.
Context, memory, and docs, all in the same surface your AI connects to. The knowledge layer the codebase can't hold on its own.
Context
Short, structured context blocks carrying the decisions, constraints, and conventions for the project. Your AI reads them on every session via MCP, so it starts informed instead of guessing.
Memory
What was tried, what was decided, what's off the table. Memory persists between sessions and agents, so a fresh chat picks up with everything the last one knew. No starting from scratch.
Documentation
ADRs, API notes, runbooks, setup guides. Pin the ones your AI should always have; it pulls the rest on demand. And when it makes a notable decision, it writes it back, so knowledge grows instead of leaking away.
A task board can't be your AI's memory.
Linear and Jira track human work in a silo your AI can't read. Here, the context lives in the same surface your AI connects to over MCP, so it reads and writes the same knowledge you manage. The plan, the context, and the progress are one thing, not three tools you stitch together.
Two levels of what your AI knows.
Context Management is the knowledge that accumulates inside each project. Builder Profile is your team's base context, the same for every project. Together, they're everything your AI builds from.
Context Management
The context, memory, and docs that build up inside a single project as you and your AI work on it.
Builder Profile
Your stack, agent, conventions, and sizing. Set once, applied to every scope you ever generate.
See Builder Profile →Give your AI a memory.
Manage the context, memory, and docs your AI builds from, in the same workspace it connects to.