lars-3010/Jarvis-Assistant
If you are the rightful owner of Jarvis-Assistant and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to dayong@mcphub.com.
Jarvis Assistant is an MCP server that provides semantic search and graph analysis tools for AI systems to intelligently discover knowledge in Obsidian vaults.
Jarvis Assistant
MCP server providing semantic search and graph analysis tools for AI systems to intelligently discover knowledge in Obsidian vaults.
Key Features
- Semantic Search: Find related content through meaning, not just keywords
- Graph Relationships: Discover connections between notes and concepts
- Claude Desktop Integration: Direct MCP protocol integration with AI tools
- Local-First: All processing happens locally for privacy and control
- Production Ready: 8 working MCP tools with robust error handling
Prerequisites
- Python 3.11+
- UV package manager
- Obsidian vault
- Claude Desktop (for MCP integration)
Quick Start
1. Install UV Package Manager
curl -LsSf https://astral.sh/uv/install.sh | sh
2. Install Dependencies
git clone <repository-url>
cd jarvis-assistant
uv sync
3. Index Your Vault
uv run jarvis index --vault /path/to/your/obsidian/vault
4. Start MCP Server
uv run jarvis mcp --vault /path/to/vault --watch
Basic Usage Example
Once running, the MCP server provides these tools to Claude Desktop & other AI systems:
# Available MCP tools (8 production-ready tools):
- search-semantic # Find related content: "productivity techniques"
- search-vault # Keyword search: "specific phrase"
- search-graph # Relationship discovery: find connected notes
- search-combined # Hybrid search combining all strategies
- read-note # Read specific files with metadata
- list-vaults # Vault management and statistics
- health-status # System health monitoring
- performance-metrics # Performance analytics and optimization
Example interaction in Claude Desktop:
"Find notes about machine learning algorithms"
Uses semantic search to find related concepts like neural networks, deep learning, AI, even if those exact words aren't in the query
Claude Desktop Integration
Add to your Claude Desktop configuration (~/.claude.json):
{
"mcpServers": {
"jarvis": {
"command": "/path/to/jarvis-assistant/.venv/bin/jarvis-mcp-stdio",
"args": ["/path/to/your/obsidian/vault"],
"type": "stdio",
"cwd": "/path/to/jarvis-assistant"
}
}
}
Replace paths with your actual directories
Architecture Overview
Claude Desktop → MCP Server → Search Services → Databases
├── Vector Search (DuckDB)
├── Graph Search (Neo4j)
└── Vault Access (File System)
Common Issues
- Binary not found: Run
uv syncto create the virtual environment - Permission denied: On Unix:
chmod +x .venv/bin/jarvis-mcp-stdio - Vault not found: Use absolute paths only
- Server won't start: Test with
uv run jarvis --help
Documentation
📚 Complete Documentation:
Quick Links
- - What is Jarvis Assistant?
- - 5-minute setup
- - Typical usage patterns
- - Complete MCP tool documentation
- - Common issues and solutions
For Different Audiences
- New Users: Start with
- Developers: See
- AI Tools: Reference
Contributing
- Read the
- Follow the
- Add tests for new functionality
- Update documentation for changes
License
MIT License - see LICENSE file for details.
Status: Production ready with 8 working MCP tools
Next: Enhanced performance and additional search capabilities
Support: See or open an issue