Inspiration

I wanted to show how multiple AI agents can collaborate like a real research team — each with a unique role.

What it does

Scout AI sends your topic through 4 agents sequentially:

  • 🧭 Strategist plans the research
  • 🔬 Analyst extracts deep facts
  • ⚔️ Critic challenges the analysis
  • 💡 Synthesizer writes the final report

How I built it

Built with Gemini 2.0 Flash API, HTML, CSS, and JavaScript. Each agent receives the previous agent's output as context — creating a true chained multi-agent pipeline.

Challenges

Getting agents to build on each other's context correctly and designing prompts that make each agent think differently.

What I learned

How to design multi-agent pipelines where each agent has a specialized role and chained context.

What's next

Add web search capability and export reports as PDF.

Share this project:

Updates