
MCP Brain Manager
by mikeybeez
Intelligent context management layer for AI knowledge systems — semantic intent routing, human-in-the-loop update proposals, and seamless multi-project stack sw
What it does
MCP Brain Manager adds an intelligent orchestration layer on top of Brain-based knowledge management systems. It lets AI assistants classify user intent semantically, load the right project context automatically at session start, and propose structured updates (progress, decisions, research notes) that require explicit human confirmation before being committed — keeping the human firmly in control of what gets written to long-term memory.
Tools
manager_init— Initialize a brain session; auto-loads context and detects continuation modepropose_update— Create a reviewable change proposal (progress, decision, or research note) with a unique IDconfirm_update— Apply a pending proposal after human review; supports inline modificationsswitch_project— Switch active project with full context preservation; can create from a templatereturn_to_previous— Pop the previous project off the stack for quick back-navigationgenerate_dashboard— Generate an Obsidian-compatible markdown dashboard for any projectsemantic_classify— Classify a message's intent (mode + confidence + reasoning)analyze_patterns— Surface work patterns and productivity insights over a configurable timeframe
Installation
Clone and build locally:
git clone https://github.com/MikeyBeez/mcp-brain-manager
cd mcp-brain-manager && npm install && npm run build
Add to your MCP config:
{
"mcpServers": {
"brain-manager": {
"command": "node",
"args": ["/path/to/mcp-brain-manager/dist/index.js"]
}
}
}
Supported hosts
Any MCP client that supports stdio transport. Confirmed via generic mcpServers config in README.
Quick install
git clone https://github.com/MikeyBeez/mcp-brain-manager && cd mcp-brain-manager && npm install && npm run buildInformation
- Pricing
- free
- Published
- 4/7/2026
- Updated
- 2 months ago
- stars