databento

This MCP server provides access to the DataBento API through Claude Desktop MCP clients. It exposes a suite of 18 tools organized into categories such as Original, Timeseries, Symbology, Metadata, Batch, and Reference, enabling real-time quotes, historical data streaming, symbol resolution, and metadata discovery all via MCP requests. You can connect to the server from Claude Desktop or other MCP clients by configuring an MCP server entry named databento and pointing it to the running Node MCP server. The server uses a DataBento API key to authenticate requests and supports the same core capabilities in both MCP mode and Claude Code Skills mode.

Once running, you can invoke tools like get_futures_quote for ES/NQ prices, get_session_info to detect the current market session, get_historical_bars for OHLCV data, timeseries_get_range for streaming historical data with flexible schemas, symbology_resolve to map symbols to instrument IDs, and a variety of metadata and batch tools to explore datasets, schemas, and perform bulk downloads. The tools are designed to be consumed as JSON requests and return structured JSON responses with data and metadata.

Prerequisites:

• Node.js v18+ (LTS) or compatible runtime
• A DataBento API key (you can obtain one from databento.com)
• Dependency installation permissions on your machine

Installation steps:

1. Clone the repository containing the MCP server:

cd ~/Dev
git clone <your-repo-url> databento-mcp-server
cd databento-mcp-server

2. Install dependencies:

npm install

3. Create and populate your environment file with the API key:

cp .env.example .env

Edit the .env file (or you can set the key directly in the MCP config env):

DATABENTO_API_KEY=db-your-api-key-here
DATABENTO_DATASET=GLBX.MDP3

4. Build the MCP server (for MCP deployment):

npm run build:mcp

5. Run the MCP server (example using node on the built output):

node dist/mcp/mcp/index.js

6. Alternatively, you can configure and start via npx (if published to npm):

# In your Claude Desktop MCP config, you can reference npx directly

7. Verify the server appears in Claude Desktop MCP configuration under the databento entry and ensure your API key is set in the environment:

{
  "mcpServers": {
    "databento": {
      "command": "node",
      "args": ["/path/to/your/dist/mcp/mcp/index.js"],
      "env": {
        "DATABENTO_API_KEY": "db-your-api-key-here"
      }
    }
  }
}

Prerequisites and notes:

• Ensure your API key is valid and has access to the DataBento datasets you plan to query.
• If you choose to publish to npm and use npx, replace the command/args accordingly as shown in the repository examples.

Environment variables: DATABENTO_API_KEY is required. Optional DATABENTO_DATASET can be set to a default CME dataset like GLBX.MDP3. If you see authentication errors, verify that the API key is active and correctly set in your environment where the MCP server runs. The MCP server supports 18 tools; consult the Available Tools section of the README for the exact names and input/output schemas. For production deployments, consider wrapping the Node process with a process manager (e.g., pm2) and securing the API key in a secret store rather than a plaintext env file. If you update to a newer version, rebuild the MCP assets using npm run build:mcp before restarting the server.