GBrain connects QM to the team's memory and connectors

QM is one AI for the whole company

Multiplayer, in Slack and on the web

Everyone gets a workspace of their own, and the team works with QM together.
Memory for each person and each room

Every person and every room keeps its own memory, files and permissions.
Not tied to one AI company

Pi, OpenCode, Codex or Claude Code can drive QM, on the model the team picks.

From QM's README on GitHub.

QM starts from what the team already knows

What one AI saves, every other AI the team uses can read.

A workspace's memory in GBrain. Each note is a file the team owns.

QM can use every connector GBrain offers

Connect an account once in GBrain. The AI never holds the key.

See every connector

The same memory and connectors in every AI the team uses

ChatGPT

QM

Claude Code

GBrain

one memory

The team's memory

Connectors

Cursor

OpenClaw

The terminal

QM on the dark line, and a few of the AIs that connect the same way. Each one opens its own page.

See every AI GBrain works with

QM connects with one admin request

One address for every AI. The team decides what each one may do.

In the GBrain workspace, make an access token for QM. GBrain shows it once.

A QM admin registers GBrain as one of QM’s MCP servers, through QM’s admin API.

PUT /v1/admin/mcp-servers/gbrain

{
  "name": "GBrain",
  "url": "https://gbrain.io/mcp",
  "auth": "bearer",
  "bearerToken": "<token>",
  "readOnly": false,
  "enabled": true
}

Every agent in the deployment can then reach the workspace. readOnly is false because GBrain can write, to memory and to the accounts connected to it.

Start a GBrain workspace

Sign in and name the workspace. Then connect QM with the steps above.

Sign in