Multi-agent system for marketing attribution
“Finds 10-15 high-quality events per query with 85%+ relevance in 60-90 seconds; saves 5+ hours per search; open-sourced on GitHub”
solo-developer
Why they built it
To solve inefficient event discovery where searches return outdated/cluttered results and CFP/sponsorship info is hard to find, saving developers hours of manual research
What worked
Multi-agent architecture for reliability; real-time WebSocket streaming for engagement; BrightData MCP for scalable scraping; form-based input for accessibility
What broke or was painful
Rate limiting and reliability in web scraping; inconsistent event website layouts affecting extraction accuracy; need for real-time feedback during processing (solved with BrightData MCP and specialized prompts)
The result
Finds 10-15 high-quality events per query with 85%+ relevance in 60-90 seconds; saves 5+ hours per search; open-sourced on GitHub