Braingent

Integrations

Codex Integration

Wire Codex / Codex CLI to your Braingent memory repo via AGENTS.md.

Codex (and Codex CLI) auto-reads AGENTS.md from the working directory. The integration mirrors Claude’s exactly — same contract, different filename.

Where the file goes

<repo-root>/AGENTS.md for project scope, or copy the same file into each working repo for portability.

If you have a Braingent memory repo at ~/Documents/repos/braingent, the file is already there as AGENTS.md — just copy it into your working repos:

cp ~/Documents/repos/braingent/AGENTS.md ./AGENTS.md

What goes in it

AGENTS.md carries the same Braingent contract as CLAUDE.md:

  • Pointer to the memory repo path.
  • Search-before-plan rules.
  • Capture-after-work triggers.
  • Privacy boundaries.

The only meaningful difference is voice — AGENTS.md typically uses plural “agents” framing because Codex is often invoked alongside other tools.

MCP setup

Codex CLI supports MCP. Configuration is similar to Claude:

{
  "mcpServers": {
    "braingent": {
      "command": "braingent",
      "args": ["mcp", "serve", "--path", "/Users/you/Documents/repos/braingent"]
    }
  }
}

Add this to Codex’s MCP config (path varies by version — see Codex CLI docs).

Smoke test

Q: What's the path to my Braingent memory repo, and what's the first
   thing you'll do before planning a non-trivial task?

Codex should answer with the absolute path and restate the contract.

Working alongside Claude

Codex and Claude can both work on the same memory repo at the same time. They’ll naturally pick up each other’s records because they read from the same Markdown.

When the same task is in flight under both, use a live multi-agent task — tasks/active/BGT-NNNN.md — for coordination.

Where to go next