A new open-source project automates competitive intelligence monitoring using CocoIndex and Tavily Search. The system continuously scans the web for competitor news, extracting structured events like product launches, funding rounds, and executive hires. It processes this data through a GPT-4o-mini extraction layer to create organized PostgreSQL databases, eliminating the need for manual tracking.

The pipeline architecture flows from Tavily's AI-native search directly into LLM extraction, producing `CompetitiveEvent` objects. These objects are stored in dual indexes for raw articles and normalized events. This approach avoids brittle scraping and turns noise into signal, allowing teams to query natural questions like "What has Anthropic been doing recently?" for actionable insights.

To analyze the data, the project includes built-in query handlers that enable SQL-powered retrieval. Users can search by competitor, filter by event type, or rank by significance scores. Configuration relies on environment variables for API keys and refresh intervals, supporting continuous monitoring via the CocoIndex CLI or interactive setup scripts for immediate deployment.

This method matters because it automates the tedious process of market surveillance. By combining AI search with structured LLM extraction, companies can maintain real-time awareness of strategic moves without dedicated analyst overhead. The MIT-licensed GitHub repository provides a complete blueprint for engineering teams to build their own automated intelligence gathering systems.