Devotd Github
NewAI agents

make0-official/brain-oss v0.3.1 — Shared Memory Server for Agents

make0-official/brain-oss is a versioned memory server that multiple coding agents share through MCP and a local SQLite store.

Repositorymake0-official/brain-oss ↗

make0-official/brain-oss is a versioned memory server that multiple coding agents share through MCP and a local SQLite store.

GitHub description: AI memory for Claude Code, Cursor, Codex, Windsurf, and every MCP agent: one shared, versioned memory in a local MCP server. Offline, no API key.

stars
★ 5
Forks
0
Language
TypeScript
License
MIT
Latest release
v0.3.1
Read original source ↗View on GitHub ↗

make0-official/brain-oss is a local memory server where multiple coding agents such as Claude Code, Cursor, Codex, and Windsurf read and write shared memory. The latest release is v0.3.1. The provided materials do not state what was newly added or changed in this version. The project is set up so that agents share a single memory store through an MCP server. Data is stored in a local SQLite file, and change history is preserved along with versions, according to the description.

The core of this setup is that instead of keeping memory separate for each agent, it provides a store used jointly by multiple tools. The README states that Make0 Brain supports importing sessions from different agents, as well as searching, classifying, and exporting memories. The installation process backs up configuration files. There is also a note that after the model is downloaded it runs offline and requires no API key. The implementation language is TypeScript and the license is MIT. It is therefore a project focused on handling the memories of several coding agents in one place.

However, the specific changes in v0.3.1 cannot be confirmed from the provided materials alone. Nor are there benchmarks in the materials showing search quality or performance when several agents use it together. Offline operation requires downloading the model in advance. So the shared store and the offline path can be considered within the feature scope shown in the documentation, but the actual effect and performance are hard to judge from these materials alone.