Stream Real-Time Stock News in Python with finlight

Stream Real-Time Stock News in Python with finlight

Polling an API for news means you're always a little behind. For live dashboards and trading signals, you want each article the moment it's published. finlight's WebSocket does exactly that, and the Python SDK makes it a few lines. In this tutorial you'll build a live stream of stock news, filtered by ticker and already scored for sentiment.

Install the SDK

pip install finlight-client

The full example

import asyncio

from finlight_client import FinlightApi, ApiConfig
from finlight_client.models import GetArticlesWebSocketParams

client = FinlightApi(config=ApiConfig(api_key="YOUR_API_KEY"))


def on_article(article):
    print(f"[{article.sentiment}] {article.title} - {article.source}")


async def main():
    await client.websocket.connect(
        request_payload=GetArticlesWebSocketParams(
            tickers=["AAPL", "NVDA"],
            includeEntities=True,
        ),
        on_article=on_article,
    )


if __name__ == "__main__":
    asyncio.run(main())

Run it, and every time a matching article is published, your on_article callback fires with the article, including its sentiment, title, source, and (because includeEntities is on) the tagged companies.

What's happening

• client.websocket.connect(...) opens a persistent connection and streams matching articles. It's an async call, so it runs inside asyncio.run(...).
• GetArticlesWebSocketParams takes the same kinds of filters as the REST API: tickers, query, sources, countries, categories, language, includeEntities, and more.
• on_article is your handler. This is where you'd push to a queue, update a dashboard, or evaluate a signal instead of just printing.

The SDK manages the connection for you, including automatic reconnection with backoff, so a dropped connection won't stop your stream.

Enhanced vs raw

client.websocket is the enhanced stream: articles arrive with full enrichment (sentiment and entities). If you only need lightweight, basic article data as fast as possible, there's also client.raw_websocket with the same connect(...) shape. Use enhanced when you want the analysis, raw when you want minimal latency.

Filter to exactly what you need

Because the WebSocket accepts the same filters as REST, you can stream a precise slice: only crypto news, only a watchlist of tickers, only certain countries. See Advanced Filtering for the full filter set.

Where to go next

• REST vs WebSocket vs Webhooks vs MCP to decide when streaming is the right choice.
• Full documentation for the WebSocket guides.

Get your free API key →