OzBrain

Context scattered across separate chat windows is the problem OzBrain is meant to solve. Operated by Monsef Holdings Pty Ltd, it works as a common brain for teams whose work runs through more than one AI agent: one structured source of knowledge that agents and teammates all read and write.

Knowledge in nested pieces

Knowledge is stored in structured articles nested inside one another, each carrying links, provenance and a freshness marker. An agent pulls only the slice it needs, such as an email body rather than a whole launch plan. OzBrain argues that loading less makes answers faster and cheaper and reduces the invention of details from context the agent never needed.

Old ideas replaced, not piled up

When newer thinking arrives, OzBrain goes back through the knowledge itself, marks the old material as replaced and points to the latest version. No agent should then answer from a decision the team has moved past.

A record of every change

Every edit is logged with the agent responsible, the content altered and the reasoning behind it. Edits are put forward before they are applied, which lets multiple agents work at the same time without overwriting one another.

Connecting and leaving

OzBrain connects through the connector menus that Claude and ChatGPT already show, and works with Claude Code, Cursor and Gemini, among others. The company says content is encrypted at rest, that it never trains on it or sells it, and that all of it can be taken out as Markdown at any time, even once an account is closed. Plans start with a free tier, continue through two paid levels, and Enterprise is arranged on request. The company says it built the tool for itself first, and its domain moved to ozbrain.com from an older address.