Lambda Capture MCP Server

What is Lambda Capture MCP Server about?

Lambda Capture MCP Server implements a Model Context Protocol (MCP) endpoint that exposes a semantic‑search API for macroeconomic data. It allows AI agents, large language models, and quant research pipelines to query economic concepts (e.g., inflation expectations) and receive the most relevant data points.

How to use Lambda Capture MCP Server?

• Remote usage: Send a JSON‑RPC request to https://mcp.lambda-capture.com/v1/mcp/ using POST. Include your Lambda Capture API token in the Authorization header. The request can be made via curl, OpenAI client tools, or any HTTP library.
• Local usage: Clone the repository, install dependencies (npm install for Node or a virtual environment for Python), then run the server locally. Configure the environment variable LAMBDA_CAPTURE_API_KEY with your API key.
• MCP client configuration: Add an entry to claude_desktop_config.json (or any MCP‑compatible client) pointing to the remote URL or the local executable.

Key features

• Semantic search: Natural‑language queries return the most relevant macro data records.
• Model‑agnostic MCP endpoint: Works with OpenAI, Claude, or any MCP‑compatible tool.
• Streaming responses: Supports text/event-stream for incremental output.
• Configurable context window: Adjust maxTokens/max_tokens to fit the model’s context size.
• Dual implementation: Both Node.js (TypeScript) and Python versions are provided.

Use cases

• Quant researchers building AI‑driven investment strategies that need up‑to‑date macro indicators.
• LLM agents performing economic analysis or generating reports based on current data.
• Automated dashboards that retrieve and display macro trends on demand.
• Academic projects exploring semantic retrieval over large economic datasets.

FAQ

Q: Do I need a Lambda Capture account? A: Yes, an API key from Lambda Capture is required for both remote and local deployments.

Q: Which models can call this MCP server? A: Any model that supports MCP, including OpenAI's GPT‑4.1, Anthropic Claude, and other compatible LLMs.

Q: How many results can I request? A: The max_results parameter controls the number of returned items; typical defaults are 3‑10.

Q: Can I run the server behind a firewall? A: Yes, the local version can be hosted on any network; just ensure the MCP client can reach the URL.

Q: How do I change the token limit? A: Edit the maxTokens (TypeScript) or max_tokens (Python) variable in the source before building.

Lambda Capture MCP Server

MCP implementation of our standard Semantic Search API for Macroeconomic Data

Remote MCP Server (streamable HTTP)

Check server status HERE

OpenAI Responses API

from openai import OpenAI

client = OpenAI()

resp = client.responses.create(
    model="gpt-4.1",
    input="Key shifts in inflation expectations",
    tools=[
        {
            "type": "mcp",
            "server_label": "lambda-capture",
            "server_url": "https://mcp.lambda-capture.com/v1/mcp/",
            "headers": {
                "Authorization": "Bearer YOUR_ACCESS_TOKEN"
            }
        }
    ]
)

print(resp.output_text)

Curl

curl -X POST "https://mcp.lambda-capture.com/v1/mcp/" \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-H "Authorization: Bearer YOUR_ACCESS_TOKEN" \
-d '{
  "jsonrpc": "2.0",
  "method": "tools/call",
  "id": 1,
  "params": {
    "name": "macroecon_semantic_search",
    "arguments": {
      "query_text": "inflation expectations",
      "max_results": 3
    }
  }
}'

curl -X POST "https://mcp.lambda-capture.com/v1/mcp/" \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-H "Authorization: Bearer YOUR_ACCESS_TOKEN" \
-d '{
    "jsonrpc": "2.0",
    "id": "1",
    "method": "list_tools",
    "params": {}
}'

Configure your MCP Client (Claude Desktop App)

Go to Claude -> Settings -> Developer -> Edit Config. Add the following to your claude_desktop_config.json

Node:

{
  "mcpServers": {
    "lambda-capture-mcp": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://mcp.lambda-capture.com/v1/mcp/",
        "--header", "Authorization: Bearer YOUR_ACCESS_TOKEN"
      ],
      "description": "RemoteMCP with Lambda Capture Macroeconomic Data API"
    }
  }
}

Local MCP Server

Pre-requisites

• Lambda Capture API key
• for MCP Typescript: Node.js 18+ (includes npx and npm)
• for MCP Python: Python 3.11+

Installation

1. Clone the repo

Node:

2. npm install to install the dependencies
3. npm run build to build the project

Python:

2. python -m venv .venv create virtual environment
3. source .venv/bin/activate activate virtual environment
4. pip install -r requirements.txt install the dependencies

Configure your MCP Client (Claude Desktop App)

Go to Claude -> Settings -> Developer -> Edit Config. Add the following to your claude_desktop_config.json

Node:

{
  "mcpServers": {
    "lambda-capture-mcp": {
      "command": "node",
      "args": [
        "/Absolute Path to/mcp-server/dist/index.js"
      ],
      "env": {
        "LAMBDA_CAPTURE_API_KEY": "Your API Key string"
      },
      "description": "Runs the Node MCP with Lambda Capture Macroeconomic Data API"
    }
  }
}

Python:

{
  "mcpServers": {
    "lambda-capture-mcp": {
      "command": "/Absolute Path to/.venv/bin/python",
      "args": [
        "/Absolute Path to/mcp-server/main.py"
      ],
      "env": {
        "LAMBDA_CAPTURE_API_KEY": "Your API Key string"
      },
      "description": "Runs the Python MCP with Lambda Capture Macroeconomic Data API"
    }
  }
}

Context Window Size

Adjust maxTokens (.ts) or max_tokens (.py) variables, based on context window size of your model (doesn't count metadata, just content tokens)

© 2025 Lambda Capture Limited (Registration Number 15845351) 52 Tabernacle Street, London, EC2A 4NJ - All rights reserved