Scan biotech news for catalyst trades with FinBERT, Alpaca, and Gemini

Description

Automatically scan major financial newswires for biotech catalyst events, score them with AI sentiment analysis, and surface ranked trade candidates — all without manual monitoring.

Who's it for
Traders and quant hobbyists who follow biotech stocks and want an automated pipeline to flag high-probability catalyst events (FDA approvals, Phase 3 readouts, M&A) before the market fully reacts.

How it works
The workflow runs four layers on a scheduled loop. Layer 0 polls PRNewswire, GlobeNewswire, and BusinessWire every minute and stores new articles in a data table. Layer 1 filters for biotech keywords, classifies the event type, and runs FinBERT sentiment analysis via HuggingFace — only PROCEED events advance. Layer 2 pulls real-time quotes and 20-day volume bars from Alpaca to calculate spread percentage and relative volume. Layer 3 sends qualifying candidates through a Gemini LLM prompt for final trade scoring and stores results.

How to set up
Add your HuggingFace API credential in n8n
Add your Google Gemini API credential in n8n
Open the Alpaca - latest bars node and paste your Alpaca API key and secret
Create five data tables: news_events, processed_events, trade_candidates, llm_analysis, market_snapshots
Activate each schedule trigger

Requirements
n8n with Data Tables enabled (self-hosted or cloud)
HuggingFace account (free tier)
Alpaca Markets account (free tier)
Google Gemini API key

How to customize
Swap the GlobeNewswire RSS feed URL to target a different sector. Adjust the biotech keyword list in the Biotech Keyword Filter node. Tune the spread % and relative volume thresholds in Eligibility Check to tighten or loosen your trade criteria.