finlight Logo

Friday, July 3, 2026

finlight in Your AI Assistant: The MCP Server

Finlight - Content Manager

finlight in Your AI Assistant: The MCP Server

What if you could just ask your AI assistant for the latest market news, filtered by sentiment and company, and get sourced, up-to-date answers? That's what the finlight MCP server does. It connects Claude or ChatGPT directly to finlight, so you can research financial news in plain language with no code.

MCP server URL: https://mcp.finlight.me

What is MCP?

MCP (Model Context Protocol) is an open standard that lets AI assistants connect to external tools and data. With finlight's MCP server, your assistant can query the real news database in real time, filter by your criteria, and synthesize the results. The answers are sourced and current, not hallucinated.

Connect Claude Desktop

Setup takes about two minutes.

  1. Install the Claude Desktop app and grab your API key from the dashboard.
  2. Open Settings → MCP Servers in Claude Desktop.
  3. Add a server with the URL https://mcp.finlight.me.
  4. Claude opens a browser to the finlight authorization page. Enter your API key and click Authorize.
  5. Return to Claude. You're connected.

Connect ChatGPT

You can connect finlight to ChatGPT by creating a Custom GPT with an Action (requires ChatGPT Plus or Enterprise):

  1. Create a new Custom GPT and open the Configure → Actions section.
  2. Set authentication to OAuth, with:
    • Authorization URL: https://mcp.finlight.me/authorize
    • Token URL: https://mcp.finlight.me/token
  3. Add the API schema (the full schema is in the documentation), then save and authorize with your API key.

What you can ask

Once connected, your assistant can use all of finlight's filtering through natural language:

"Show me all negative sentiment news about NVIDIA from the last 24 hours."

"Find all news mentioning my watchlist: AAPL, MSFT, GOOGL, AMZN, sorted by most recent."

"Summarize today's coverage of the Federal Reserve meeting, grouped by topic."

"What's the sentiment distribution for Tesla news this quarter?"

It can search by keyword, filter by ticker, sentiment, source, country, and date range, pull full article content, and list available sources.

Who it's for

  • Analysts and portfolio managers monitoring tickers and sentiment without tab-switching.
  • Research teams synthesizing coverage from many sources at once.
  • Quant and trading teams doing fast, exploratory research before building something.

For developers

Want to integrate the same capability into your own app? The developer guide covers direct HTTP calls, TypeScript and Python examples, the OAuth flow, and error handling.

Get started

  1. Get your API key (free tier available).
  2. Connect Claude or ChatGPT with the steps above.
  3. Start asking. Try the example prompts to see what's possible.

Get your free API key →