Second Brain MCP
Connect agents to a self-maintaining notes vault with search, wikilinks, and compressed long-term memory across projects.
Overview
second-brain MCP is a Build-phase MCP server that lets agents search and maintain a linked knowledge vault with compression for long-running memory.
What is this MCP server?
- Figure-level search over vault content for precise retrieval
- Auto-wikilinks to surface related notes without manual graph upkeep
- Memory compression to keep agent context usable as the vault grows
- PyPI package mcp-second-brain (0.1.0) with stdio transport
- GitHub repository second-brain-mcp for local vault wiring
- Package version 0.1.0 on PyPI identifier mcp-second-brain
- Stdio transport in published server.json
- GitHub repository id 1253048999 at ddmanyes/second-brain-mcp
Community signal: 2 GitHub stars.
What problem does it solve?
Project knowledge fragments across chats and files so agents repeat research and miss connections you already documented.
Who is it for?
Builders who already keep a linked notes vault and want agents to read, navigate, and maintain it during implementation.
Skip if: Teams wanting turnkey cloud PKM with no local vault setup, or purely transactional APIs without a notes backend.
What do I get? / Deliverables
Agents pull structured, interlinked notes from your vault with tighter recall as the corpus grows.
- Search hits at figure-level granularity from the vault
- Maintained wikilink graph and compressed memory summaries for agent sessions
Recommended MCP Servers
Journey fit
Documentation and captured decisions anchor the build phase, but a second brain pays off from early research through ongoing iteration. Docs is the canonical shelf for knowledge bases agents query while coding, planning, and writing specs.
How it compares
Local knowledge-vault MCP, not a disposable-email or generative-media tool server.
Common Questions / FAQ
Who is Second Brain MCP for?
Indie developers who maintain a personal knowledge base and want Claude Code or similar agents to search and update it via MCP.
When should I use Second Brain MCP?
Use it whenever the agent needs durable context—specs, research, architecture notes—across sessions instead of one-off paste-ins.
How do I add Second Brain MCP to my agent?
Install the PyPI package mcp-second-brain and register it as a stdio MCP server per your client docs, pointing at your configured vault.