LBrain

ai.lbrainPublished by lbrain.aiv0.1.3Updated Oct 24, 2026

Local-first agent memory. Source, date, bind/abstain. Permanent name: lbrain.ai/claim.html

The publisher's own server: BrainDog

VerifiedSTDIODesktop onlyKnowledge & MemoryDeveloper Tools

Overview

AI-generated overview

Wires coding agents to a local-first LBrain memory engine so they can recall, bind, or abstain on stored notes.

What you can do
This package installs and wires hosts to the local LBrain memory engine's stdio MCP server, so an assistant can query a personal memory store. It detects supported coding agents and writes the lbrain MCP entry into their configs, copies a companion skill into each host's skills directory, and can inject a session-start recall call. A Python wrapper lets OpenAI-style clients recall first and abstain when nothing binds, with writes opt-in.
When to use it
Use it when you keep notes or documents locally and want coding agents to recall from that store across sessions, with a bias toward abstaining rather than guessing. It suits users who already run the LBrain engine and want several agent hosts wired consistently.
Requirements
A local Python environment: install the LBrain engine with pip install "lbrain[local]", initialize it with a source directory, import and embed, then install this package and run its installer. The MCP server runs locally over stdio; no remote endpoint, account, API key, or environment variable is declared.
Before you install
The installer writes to agent configuration files, though it merges only the lbrain key, leaves other MCP servers alone, and backs up a config once before the first write. It does not set LBRAIN_HOME unless asked. The engine's HTTP transport has no authentication, so it should not be exposed to a network; this installer enables stdio only.

Installation

In SourceWeft

  1. Open LBrain in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

LBrain coding-agents

Install LBrain where the agent already lives.

The free name is generic. Make it permanent: https://lbrain.ai/claim.html

Native MCP + companion skill + fail-closed session-start for Claude Code, Codex, Cursor, Copilot, Grok Build, Gemini CLI, Antigravity, and OpenClaw.

This package does not modify the LBrain engine. It wires hosts to lbrain mcp. The Official MCP Registry / VS Code @mcp gallery entry also execs that same stdio server.

VS Code: search LBrain in Extensions, or @mcp lbrain. Marketplace: https://marketplace.visualstudio.com/items?itemName=metavolve-labs.lbrain

Install the engine first

bash
pip install "lbrain[local]"lbrain init --source ~/noteslbrain import && lbrain embed --stale

Wire your coding agents

bash
pip install lbrain-coding-agentspython3 -m lbrain_agents statuspython3 -m lbrain_agents install            # lists detections, changes nothingpython3 -m lbrain_agents install all        # every detected harnesspython3 -m lbrain_agents install claude-code grok-buildpython3 -m lbrain_agents install all --dry-run

A bare install is a no-op. That is deliberate: wiring every agent on the machine never happens by accident.

Paste to a desktop AI (no terminal)

Some people never open a shell. They paste this into Claude, Cursor, Codex, Copilot, or Grok and the agent does the work:

Install LBrain as my local memory.
pip install "lbrain[local]" lbrain-coding-agentsIf python3 -m lbrain_agents works here, run: python3 -m lbrain_agents install allOtherwise add an MCP server named lbrain: command lbrain, args ["mcp"], stdio.
Then call whoami. Prefer binds. Near-miss is not an answer. Abstain if nothing binds. SUPERSEDED must not govern. Leave LBRAIN_HOME alone if it is already set.
https://lbrain.ai/integrations.html

Existing lbrain MCP entries are left untouched unless you pass --force. Skills are copied to each host's skills directory. Session-start injects lbrain whoami (Claude Code).

Two-line wrap

python
from openai import OpenAIfrom lbrain_agents.wrap import wrap_openai
client = wrap_openai(OpenAI(), home="~/.lbrain")

Recall first. If nothing binds, the model is told to abstain. Writes are opt-in (remember=True).

Agent Plugin

lbrain_agents/share/agent-plugin/ is a portable bundle (plugin.json + mcp.json + skill) for hosts that speak the Agent Plugins spec.

Cookbooks

  1. Abstain when the record does not bind
  2. SUPERSEDED does not govern
  3. Model swap, same mind

See lbrain_agents/share/cookbooks/.

Safety

  • Merges only the lbrain key. Other MCP servers stay.
  • Backs up a config once as *.lbrain-backup before the first write.
  • Does not set LBRAIN_HOME unless you pass --home. A working seat mount is not overwritten.
  • HTTP MCP is not enabled by this installer. stdio only. The engine's HTTP transport has no auth; do not bind it to the world.

What this package is not

This is the free installer and the serve-boundary skill. It is not LBrain Connect, Managed LBrain, or LBrain Govern. It does not issue gcx:// names, inscribe permanence, or ship study gold sets. The engine's resolver client is open; registrar issuance is not in this tree.

License

BSD-3-Clause. Metavolve Labs, Inc.

Patents pending. The licence covers the code; it does not grant patent rights.

Source: README.md at commit 3ca1a1c

Tools

0
SourceWeft has not started this package in its sandbox yet, so its tools are not known.

Version history

1
  1. v0.1.3LatestSep 16, 2026

Score

50% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts

Other Brain integrations

BrainDog

dog.brainOfficial publisher

BrainDog is a hosted record of projects, decisions and money that connected AI agents read before starting work and write their results into.

Streamable HTTPWeb executableAuth
Productivity & WorkflowCredential requiredNov 5, 2026

Codebase Brain

Published by GitHub account Nomannazir

Indexes a local or public git codebase so an assistant can search it, trace git history, check change impact and find stale docs, all with path:line citations.

STDIODesktop only
Web Search & ScrapingNot validated yetNov 7, 2026

Personal Brain

Published by GitHub account zhang-liz

Gives an AI assistant memory of your professional network, stored as a graph in FalkorDB, so it can look up contacts, find people by topic, and save who you…

STDIODesktop only
Databases3 toolsNov 6, 2026

Second Brain

Published by GitHub account ara-5

Shared Graph-RAG memory for AI agents: markdown vault, wikilink recall, tasks, messaging.

STDIODesktop only
Knowledge & Memory12 toolsNov 7, 2026

Make0 Brain

Published by GitHub account make0-official

A local, offline memory server that lets Claude Code, Cursor, Codex and other MCP agents share one versioned store of durable facts.

STDIODesktop only
Knowledge & MemoryNot validated yetNov 7, 2026