ScoutAI - AI location scouting.

Inspiration

Film location scouting is much more than finding a place that looks right. A location also needs to fit the story, mood, budget, access requirements, permits, and practical realities of production.

I wanted to build an AI assistant that could take a filmmaker's creative idea and turn it into meaningful, production-ready location research.

That's how ScoutAI started.

What it does

ScoutAI lets filmmakers describe their scene naturally — including the mood, setting, time period, budget, and production requirements.

It then:

• Understands the creative brief
• Generates targeted searches
• Researches real-world locations
• Evaluates and scores candidates
• Investigates access and permit considerations
• Provides supporting sources
• Generates a structured scout report and producer-ready memo

Instead of simply asking "Where should I film?", ScoutAI helps answer "Could I actually film this scene here?"

How I built it

ScoutAI is built with Next.js, TypeScript, Gemini, Google Cloud Vertex AI, Parallel Search, and Supabase.

The scouting workflow runs as a staged agentic pipeline: understanding the brief, searching, evaluating candidates, verifying details, and assembling the final scouting packet.

Production authentication uses Google Cloud Workload Identity Federation, avoiding downloadable service-account keys.

Challenges & Learnings

The biggest challenge was going beyond visual recommendations. A location can look perfect but still be difficult or expensive to shoot.

I learned that useful agentic AI needs a clear objective and a structured workflow. ScoutAI combines creative understanding with real-world research and makes the agent's progress visible to the filmmaker.

What's next

ScoutAI could grow into a complete production-location assistant, helping teams move from an initial creative brief to researched, shareable location options much faster.

Built With

• nextjs
• react
• tailwindcss
• typescript