Lambda Capture MCP Server: Fed BoE ECB Forecasts

Overview

The Lambda Capture MCP Server is a Model Context Protocol (MCP) implementation that exposes curated macroeconomic forecasts and narrative content from the Federal Reserve, Bank of England and European Central Bank. It packages semantic search and retrieval tools as an MCP-compatible HTTP endpoint so LLMs and agent frameworks can request targeted macroeconomic context during generation or tool execution.

You can run the server locally (Node or Python) or use the hosted endpoint (https://mcp.lambda-capture.com/v1/mcp/). It is designed to be used as a remote MCP tool for the OpenAI Responses API, Anthropic-style MCP integrations, or desktop MCP runners (e.g., Claude Desktop). The service returns structured search results (ranked excerpts, metadata and source links) and supports streaming responses for low-latency agent interactions.

Features

• Semantic search over macroeconomic commentary, forecasts and tables from Fed / BoE / ECB
• MCP-compatible JSON-RPC HTTP endpoint (supports streaming via SSE)
• Remote hosted server and runnable local server (Node.js or Python)
• Simple authentication via API key header
• Tools for listing available MCP tools and performing ranked semantic retrievals
• Configurable context token budget to match model context windows

Installation / Configuration

Prerequisites:

• Lambda Capture API key (account required)
• Node.js 18+ for the TypeScript/Node server
• Python 3.11+ for the Python server

Clone the repo:

git clone https://github.com/lambda-capture/mcp-server.git
cd mcp-server

Node (build & run):

npm install
npm run build
# Run the built server (example)
LAMBDA_CAPTURE_API_KEY="your_api_key" node dist/index.js

Python (venv & run):

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
# Run
export LAMBDA_CAPTURE_API_KEY="your_api_key"
python main.py

Remote server usage (hosted):

• Endpoint: https://mcp.lambda-capture.com/v1/mcp/
• Provide Authorization: Bearer YOUR_ACCESS_TOKEN header
• To stream responses, set Accept: text/event-stream

Configure Claude Desktop (example snippets to run a local MCP adapter):

Node entry in claude_desktop_config.json:

{
  "mcpServers": {
    "lambda-capture-mcp": {
      "command": "node",
      "args": ["/absolute/path/to/mcp-server/dist/index.js"],
      "env": { "LAMBDA_CAPTURE_API_KEY": "YOUR_API_KEY" },
      "description": "Lambda Capture MCP (Node)"
    }
  }
}

Python entry:

{
  "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" },
      "description": "Lambda Capture MCP (Python)"
    }
  }
}

Note: adjust the server’s maxTokens / max_tokens setting to match your target model’s context window (these variables control how many content tokens are returned).

Available Tools / Resources

Common MCP tool names and parameters exposed by the server:

Example call parameters for macroecon_semantic_search:

• query_text: free-text query (e.g., “inflation expectations”)
• max_results: number of top matches to return

Examples

OpenAI Responses API (Python) — calling MCP as a tool:

from openai import OpenAI
client = OpenAI()

resp = client.responses.create(
    model="gpt-4.1",
    input="Summarize recent 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)

Direct HTTP (curl) — semantic search:

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
      }
    }
  }'

List available tools:

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

Use Cases

• Augment LLM-generated macroeconomic commentary with sourced Fed / BoE / ECB passages to improve factuality and traceability.
• Build a research assistant that retrieves forecast excerpts and times series context for policy decisions or academic work.
• Create a quant workflow that semantically searches policy statements and central bank forecasts to generate trading signals or risk indicators.
• Power dashboards that surface the most-relevant central bank commentary when analysts query topics like “wage growth” or “market-based inflation expectations.”

Links & Resources

• Repository: https://github.com/lambda-capture/mcp-server
• Hosted endpoint health/status: https://mcp.lambda-capture.com/
• Semantic Search API reference (implementation basis): Lambda Capture Semantic Search API docs (in repo)

For best results, match the MCP server’s max_tokens configuration to your model’s available context size and authenticate requests with a valid Lambda Capture API key.