OzBrain for Claude Code
One brain behind every repo.
Updated
OzBrain connects to Claude Code with one MCP command at https://ozbrain.co/api/mcp. After you approve once, the agent reads your routing index, pulls the articles a task needs, and stages what it learns so Claude on the web, ChatGPT, and Cursor share the same facts. Your CLAUDE.md keeps the build commands; the brain keeps you.
What the agent learns mid-build stops dying with the session, and the next repo starts with everything the last one taught it. Grant the brain to a teammate and their agents read it too.
One command, one approve, then the tools show up.
- 01
Add the connector
Run: claude mcp add --transport http ozbrain https://ozbrain.co/api/mcp
- 02
Sign in
Run /mcp inside Claude Code, choose ozbrain, and approve in the browser window that opens. Once.
- 03
Work
The brain’s tools appear to the agent. At the start of a task it reads the routing index and pulls the articles the work needs.
What Claude Code can do once connected
The connector exposes your brain as tools. The agent decides when to use them, and every write is attributed in the article history.
Read the routing index
A compact listing of every article and what it covers, so the agent finds knowledge by structure instead of whatever made it into the repo’s context file.
Start any repo with full context
Your stack and your conventions, your clients and your decisions: available in a fresh clone with nothing copied over.
Stage what it learns
Decisions and gotchas worth keeping get staged and routed mid-build, so they exist after the session ends and in every other agent.
Surface conflicts before they land
If a write contradicts what the brain holds, the write pauses and the agent sees the conflict. It can record an exception; a different agent or you must approve a supersede.
Your CLAUDE.md gets thinner
Keep it for build commands, generated directories, and repo quirks. Move everything you have been copying between repos into the brain, and the file goes back to doing the one job it is good at.
Keep Claude Code. Connect the brain under it.
Repo files stay for repo quirks. OzBrain holds the knowledge that repeats across repos, agents, and people you grant access to.
Create your brain, run the MCP add command, approve once, then ask Claude Code to read the routing index.
Questions people ask before connecting
Does this replace CLAUDE.md?
No, it shrinks it. Repo-specific instructions stay in the file; knowledge that repeats across repos moves to the brain. The full argument is at OzBrain vs CLAUDE.md files.
Which plans support it?
Any plan that includes Claude Code. The connector uses the same MCP support Claude Code ships with; there is nothing extra to install.
Is this the same brain as Claude on the web?
Yes, and that is the point. What the coding agent stages, Claude in the app can read, and the other way around. One brain, every surface, with each write attributed to the agent that made it.