What the finlight API Can Do

finlight is a financial news API. It takes the constant, messy stream of global market news and turns it into clean, structured, AI-enriched data you can query, stream, or have pushed to you. Instead of scraping websites and parsing HTML yourself, you get one consistent interface with sentiment, company tagging, and rich filtering built in.

This page is the map. It covers what finlight gives you, the four ways to consume it, and where to go next. If you just want to make your first call, start with the Quick Start guide.

What's in every article

Every article finlight returns is structured and enriched, not just a headline and a link. You get:

• Core fields: title, summary, full content (optional), source, publish date, language, and images.
• Sentiment: each article is scored positive, neutral, or negative with a confidence value, so you can quantify tone instead of reading every story.
• Company entities: finlight tags the companies a story is about and resolves them to real identifiers: ticker, ISIN, OpenFIGI, sector and industry, plus every exchange the company is listed on (primary and secondary listings, each with exchange code and country).
• Classification: each article is tagged with the countries it affects and one or more of 13 categories.

This means you can answer questions like "how did sentiment on AAPL shift this week?" or "show me every regulation story affecting a set of companies" without building an NLP pipeline yourself.

Four ways to get the data

The same enriched data is available through four delivery methods. You pick the one that fits your workload.

REST API

Query news on demand: send a request, get a page of matching articles back. Best for search, backfills, scheduled dashboards, and any request/response workflow. The advanced POST /v2/articles endpoint supports the full filter set, and there are also endpoints to fetch a single article by URL and to list available sources. See the REST docs.

WebSocket

Stream articles in real time as they are published, with no polling. It comes in two flavors: an enhanced stream with full enrichment (sentiment and entities), and a lightweight raw stream for when you only need the basics fast. Ideal for live dashboards, trading signals, and monitoring. See the WebSocket docs.

Webhooks

Have finlight push news to your own endpoint the moment it matches your filters. Enriched and raw variants are available, and deliveries are signed so you can verify them. Best for event-driven pipelines, alerts, and automations that should run without you polling. See the webhook docs.

MCP server

Connect finlight to an AI assistant like Claude or ChatGPT through the Model Context Protocol, then ask for market news in plain language. Best for research and ad-hoc analysis without writing code. See the AI assistant docs.

As a rule of thumb: REST for on-demand, WebSocket for live, webhooks for push automation, and MCP for AI-driven research. A dedicated comparison guide is coming soon.

Precise filtering

finlight is built around getting exactly the news you want. Across REST, WebSocket, and webhooks you can filter by:

• Tickers: companies like AAPL or NVDA.
• Countries: ISO country codes, to focus on specific markets.
• Categories: any of the 13, from markets and economy to crypto, regulation, and energy.
• Sources: include or exclude specific outlets, or opt in to additional ones.
• Date ranges and a full query language with boolean operators and field filters, for example (ticker:AAPL OR ticker:NVDA) AND NOT source:reuters.com.

For the full query syntax, see the advanced query docs.

Global, multilingual coverage

finlight covers news in 9 languages: English, Simplified Chinese, Hindi, Spanish, Japanese, Korean, Portuguese, Arabic, and Russian. Combined with country and category filters, that lets you track a market in its own language and region instead of relying on English-only coverage.

Built for developers

You can call finlight with plain HTTP, or use the official SDKs:

• Node / TypeScript: npm install finlight-client
• Python: pip install finlight-client

Both wrap the REST and WebSocket APIs, handle auth and retries, and return typed objects. Authentication is a single X-API-KEY header on every request, using a key you generate in the dashboard. A minimal request looks like this:

curl -X POST https://api.finlight.me/v2/articles \
  -H "X-API-KEY: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "tickers": ["AAPL"], "pageSize": 5 }'

Plans

finlight has plans from a free tier for testing up to enterprise, differing mainly in request volume, history depth, real-time access, and features such as custom sources. The free plan is REST-only and ideal for evaluation. For current limits and what each tier unlocks, see finlight.me/pricing.

Start building

• New here? Follow the Quick Start guide to make your first call in five minutes.
• Ready to go deep? Browse the full documentation.
• Questions? Join the Discord or email info@finlight.me.

Get your free API key →