Lbrain MCP server
{
"mcpServers": {
"lbrain": {
"command": "uvx",
"args": [
"lbrain-coding-agents"
]
}
}
}{
"mcpServers": {
"lbrain": {
"command": "uvx",
"args": [
"lbrain-coding-agents"
]
}
}
}{
"mcpServers": {
"lbrain": {
"command": "uvx",
"args": [
"lbrain-coding-agents"
]
}
}
}{
"servers": {
"lbrain": {
"command": "uvx",
"args": [
"lbrain-coding-agents"
]
}
}
}{
"mcpServers": {
"lbrain": {
"command": "uvx",
"args": [
"lbrain-coding-agents"
]
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsoninstall uvx lbrain-coding-agentsversion 0.1.3
What it does
LBrain coding-agents connects local-first agent memory to coding assistants like Claude Code, Cursor, Copilot, and others via MCP. It lets you wire your notes and documents as persistent memory that agents can query and update during development work.
- Query your local notes and documentation as context while coding, with the agent recalling relevant information before generating solutions.
- Install LBrain memory across multiple coding tools on your machine so they all reference the same local knowledge base.
- Add memory capabilities to Claude, Cursor, or other agents by pasting installation instructions directly into the chat interface.
Tools this server exposes
| Tool | Does |
|---|---|
| wrap_openai | Python wrapper for OpenAI client that injects LBrain memory recall before requests and enables optional memory writes. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.