Knowledge Vault

Provides a local MCP server to search your Fondu Knowledge Vault using hybrid search and reranking.
  • python

0

GitHub Stars

python

Language

6 months ago

First Indexed

3 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs
{
  "mcpServers": {
    "fondutech-mcp_tools": {
      "command": "python",
      "args": [
        "mcp_fondu_search_user_context/server.py",
        "--host",
        "127.0.0.1",
        "--port",
        "8080"
      ],
      "env": {
        "FONDU_AUTH_TOKEN": "YOUR_TOKEN"
      }
    }
  }
}

You can run a local MCP server that connects to your Fondu Knowledge Vault to perform fast, hybrid semantic and keyword searches over your personal knowledge base. This server is designed to be started locally and accessed by MCP clients to return relevant knowledge snippets with contextual metadata.

How to use

Run the server on your machine and connect your MCP client (such as Claude Desktop) to it. The server exposes endpoints for health checks, real-time updates via Server-Sent Events, and message handling, enabling you to search your vault with rich results and metadata. Use this setup to quickly locate notes, documents, and references from your personal knowledge base using a natural language query and optional keyword priorities. When you start a session, supply your authentication token through one of the supported methods so the server can securely access Fondu’s API.

How to install

Prerequisites: Python 3.8+ and access to the Fondu Knowledge Vault API with a valid authentication token.

Step 1. Create a development environment and install dependencies.

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Step 2. Configure authentication. You can provide the token directly, through environment variables, a config file, or a token file. The token is required to access Fondu’s API.

Step 3. Start the MCP server using the recommended method.

Option A: Start with the direct Python command (recommended) after activating the virtual environment.

source .venv/bin/activate
python mcp_fondu_search_user_context/server.py --host 127.0.0.1 --port 8080

Option B: Start using the provided run script as an alternative.

./run.sh

Additional configuration and usage notes

Authentication options include: explicit token passed to the tool, environment variables, config files, or token files. Ensure your token is valid and accessible by the process. If you choose environment variables, set one of these: FONDU_AUTH_TOKEN or FONDU_API_TOKEN.

Environment variables you may set include the following (example placeholders shown): FONDU_AUTH_TOKEN=your-token-here.

Health, real-time communication, and endpoints

The server provides a health endpoint to verify it is running and ready. It also offers a Server-Sent Events (SSE) endpoint for real-time MCP communication and a message handling endpoint for interaction with MCP clients.

Claude Desktop configuration

To connect Claude Desktop to the MCP server, configure the MCP server entry with the Python command and environment variable for authentication.

{
  "mcpServers": {
    "knowledge_vault": {
      "command": "python",
      "args": ["mcp_fondu_search_user_context/server.py","--host","127.0.0.1","--port","8080"],
      "env": {
        "FONDU_AUTH_TOKEN": "your-auth-token-here"
      }
    }
  }
}

Alternatively, you can run the server with the run script by setting the environment variable for authentication.

Testing and verification

Health check: verify the server is running by querying the health endpoint at http://127.0.0.1:8080/health. You should receive a responsive status indicating the server is healthy.

Tool usage example: query your knowledge vault with a natural language prompt and, if available, a set of prioritized keywords to influence keyword matching.

Security and troubleshooting

Ensure the authentication token is kept secure and not exposed in logs or shared files. If you encounter authentication errors, recheck token validity and confirm it is correctly supplied through your chosen method.

If the server cannot reach Fondu API endpoints, verify internet connectivity and API availability. Ensure no firewall blocks outbound HTTPS requests to the Fondu API.

Deployment and notes

The server is designed to run locally for development and testing. When deploying to production, ensure the Hosting environment supports Python 3.8+, Python dependencies are installed, and the port is accessible to MCP clients.

Available tools

gather_relevant_user_knowledge

Search your knowledge vault using hybrid semantic and keyword matching.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational