rwnalds/engram
by Various
Self-hosted MCP server + dashboard giving Claude Code, Cursor & Hermes agents shared long-term memory over a git-backed markdown vault. Agent-native, Obsidian-compatible second bra
MCP
rwnalds/engram
Added 13 July 2026
Overview
A self-hosted MCP server and dashboard that gives Claude Code, Cursor, and Hermes agents shared long-term memory over a git-backed markdown vault. It uses markdown RAG without a vector database and is compatible with Obsidian.
Best for
Best for
Developers who want a simple, self-hosted memory layer for multiple AI agents
Use cases
- Give Claude Code persistent memory across sessions
- Share context between Cursor and Hermes agents
- Maintain a git-versioned knowledge base for agents
How to use
Install
bun install && bun dev Tools exposed
bun
Tested with
Claude Code, Codex, Hermes, openclaw, Cursor, Cline, Windsurf, Claude Desktop
Notes
A self-hosted MCP server and dashboard that gives Claude Code, Cursor, and Hermes agents shared long-term memory over a git-backed markdown vault. It uses markdown RAG without a vector database and is compatible with Obsidian.
3 stars on GitHub. Last updated 2026-07-13. Licensed MIT.
Use cases
- Give Claude Code persistent memory across sessions
- Share context between Cursor and Hermes agents
- Maintain a git-versioned knowledge base for agents
Pros
- Self-hosted gives full control over data and privacy
- No vector database reduces infrastructure complexity
- Git-backed vault provides versioning and rollback
Cons
- Low star count (3) indicates early stage or limited adoption
- Requires self-hosting and ongoing maintenance
- Markdown RAG may be less performant than vector search for large datasets
Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.
Pros
- Self-hosted gives full control over data and privacy
- No vector database reduces infrastructure complexity
- Git-backed vault provides versioning and rollback
Cons
- Low star count (3) indicates early stage or limited adoption
- Requires self-hosting and ongoing maintenance
- Markdown RAG may be less performant than vector search for large datasets
Pairs with
Other entries in the index that connect to this one. Click through to see the chain.
Get the free Developer’s Field Guide
A 27-page field guide to the AI coding workflow with Claude. Claude Code, MCP servers, the prompt patterns that work, and what to delegate. Free.
Enter your work email. We send it straight over, plus a few short notes worth knowing. Unsubscribe any time.