Provides a unified gateway with context ranking using LangGraph agents, enabling more intelligent retrieval from multiple data sources.

Enables entity extraction and knowledge graph construction using Neo4j, allowing AI agents to query relationships between concepts in your notes.

1. Click on "Install Server".

2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.

3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@memory-mcp Search my notes for project ideas"

That's it! The server will respond to your query, and you can continue using it as needed.

Here is a step-by-step guide with screenshots.

┌──────────────────────────────────────────────────────────────┐
│ Unified Gateway (:8000) │
│ FastAPI + LangGraph Agent Pipeline │
│ analyze → execute → aggregate → rank │
├──────────────────────────────────────────────────────────────┤
│ │
│ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌────────────────┐ │
│ │ Core MCP │ │Files MCP │ │ Browser │ │ Code MCP │ │
│ │ │ │ │ │ MCP │ │ │ │
│ │• Notes │ │• PDF │ │• Chrome │ │• Git commits │ │
│ │• Search │ │• DOCX │ │• Edge │ │• Repo stats │ │
│ │• Events │ │• Text │ │• Firefox │ │• VSCode recent │ │
│ │• Save │ │ │ │ │ │ │ │
│ └──────────┘ └──────────┘ └──────────┘ └────────────────┘ │
│ │
│ ┌──────────────────┐ ┌──────────────────────────────────┐ │
│ │ Conversations MCP│ │ Knowledge Graph MCP │ │
│ │ │ │ │ │
│ │• Claude exports │ │• Entity extraction (regex-based) │ │
│ │• ChatGPT exports │ │• Graph search (Neo4j) │ │
│ │• Search & parse │ │• Path finding │ │
│ └──────────────────┘ └──────────────────────────────────┘ │
│ │
├──────────────────────────────────────────────────────────────┤
│ ChromaDB (vectors) │ Neo4j (graph) │ SQLite (events) │
└──────────────────────────────────────────────────────────────┘

git clone https://github.com/Shaktisinhchavda/memory-mcp.git
cd memory-mcp
uv sync

cp .env.example .env
# Edit .env if needed — defaults work out of the box on any OS

echo "# My Project Ideas" > data/notes/ideas.md

# Index notes into ChromaDB (incremental — only new/modified files)
uv run python scripts/index_notes.py

# Full rebuild (if needed)
uv run python scripts/index_notes.py --force

# Start Neo4j (optional — for knowledge graph)
docker compose up -d

# Build knowledge graph from notes
uv run python scripts/build_graph.py

{
 "mcpServers": {
 "memory-mcp": {
 "command": "uv",
 "args": ["--directory", "/path/to/memory-mcp", "run", "python", "core_mcp/server.py"]
 },
 "files-mcp": {
 "command": "uv",
 "args": ["--directory", "/path/to/memory-mcp", "run", "python", "connectors/files_mcp/server.py"]
 },
 "browser-mcp": {
 "command": "uv",
 "args": ["--directory", "/path/to/memory-mcp", "run", "python", "connectors/browser_mcp/server.py"]
 },
 "code-mcp": {
 "command": "uv",
 "args": ["--directory", "/path/to/memory-mcp", "run", "python", "connectors/code_mcp/server.py"]
 },
 "conversations-mcp": {
 "command": "uv",
 "args": ["--directory", "/path/to/memory-mcp", "run", "python", "connectors/conversations_mcp/server.py"]
 },
 "graph-mcp": {
 "command": "uv",
 "args": ["--directory", "/path/to/memory-mcp", "run", "python", "knowledge_graph/server.py"]
 }
 }
}

uv run uvicorn gateway.server:app --host 127.0.0.1 --port 8000

memory-mcp/
├── core_mcp/ # Phase 1 — Core MCP server
│ ├── server.py # Main MCP server (stdio)
│ ├── tools/ # Notes + search + write-back tools
│ ├── vector_store/ # ChromaDB integration
│ └── event_logger/ # File watcher + SQLite
├── connectors/ # Phase 2 — Data connectors
│ ├── files_mcp/ # PDF, DOCX, text reader
│ ├── browser_mcp/ # Chrome, Edge, Firefox history
│ ├── calendar_mcp/ # Google Calendar (OAuth)
│ ├── code_mcp/ # Git + VSCode activity
│ └── conversations_mcp/ # Claude/ChatGPT exports
├── knowledge_graph/ # Phase 3 — Neo4j graph
│ ├── extractor.py # Entity extraction (regex-based)
│ ├── graph_store.py # Neo4j CRUD + search
│ └── server.py # Graph MCP server
├── gateway/ # Phase 4 — Unified gateway
│ ├── agent.py # LangGraph orchestrator
│ ├── router.py # Smart query router
│ ├── ranker.py # Multi-signal ranking
│ ├── context.py # PersonalContext model
│ └── server.py # FastAPI gateway
├── scripts/ # Utility scripts
│ ├── index_notes.py # Build vector index (incremental)
│ ├── build_graph.py # Build knowledge graph
│ └── start_watcher.py # Start file watcher
├── data/ # Your personal data (git-ignored)
│ ├── notes/ # Markdown notes
│ ├── files/ # Documents
│ └── conversations/ # Exported AI chats
├── docker-compose.yml # Neo4j container
├── pyproject.toml # Dependencies
└── .env.example # Config template

Looking for Admin?