Your workspace

Settings

Connect the models your brain thinks with, and the agents and tools that read from it.

Models

Bring your own model for each job the brain does. Keys are encrypted before they are stored.

Three jobs, one model each. An unset slot uses Main, and with nothing connected the managed model answers.

Provider

All providers

Search providers

The list Colonizer supports, from its provider catalog. Listing a service is not an endorsement.

Endpoint

API key

Needed again for every connect or change — a stored key never comes back. It is encrypted, and only the last 4 characters are ever shown.

Model

The exact model name your provider uses.

We send a few test requests first, so this can take a few seconds. Only workspace owners and admins can connect models.

A token lets a coding agent or the CLI read your brain without a browser. Revoking one stops it at once.

Your API tokens
Name	Prefix	Access	Created	Actions

Your agent reads the brain over MCP. Create a token above, then use it here.

MCP endpoint

Claude Code

Replace <token> with your token.
Any other MCP client

Add a streamable HTTP server with the endpoint above and this header:

Authorization: Bearer <token>
The livingbrain CLI

Signs in through your browser with a one-time code. No token needed.

Not available yet

Proactivity When the brain may speak up without being asked, such as a nightly tidy-up or briefs on new people.
Tools Which tools the brain may call on each person's behalf.
Automations Scheduled passes: a daily digest, a dream pass and a lint pass.
Skills A shared library of prompts your whole team can use.
Colonizer Hand bigger jobs to a colony that comes back with a pull request.

Built with

What Living Brain runs on, from the Factory Zero registry for this venture.