Inspiration
Film production involves a lot of repetitive work—breaking down scripts, researching locations, and planning scenes. We wanted to build an AI assistant that could turn a screenplay into useful, production-ready insights.
What it does
ScenePilot analyzes screenplays, extracts characters, locations, props, and scene details, and performs AI-powered research to help filmmakers plan production faster. It also allows users to access previous research.
How we built it
We built ScenePilot using Python, FastAPI, Gemini, and Google ADK. We designed multiple specialized AI agents that work together, with parallel processing to make the workflow faster. The application has a web interface for uploading and analyzing screenplays and is deployed online.
Challenges We Faced
Our biggest challenges were understanding the complex context of screenplays, coordinating multiple AI agents, handling model availability, and making the application reliable in production. Moving from a local prototype to a deployed application also required significant debugging and iteration.
What We Learned
We learned that building a useful AI product requires more than good prompts. Agent orchestration, system architecture, error handling, deployment, and user experience are equally important. We also learned how parallel multi-agent workflows can make complex AI tasks more efficient.
What's Next for ScenePilot
We plan to expand ScenePilot with shot breakdowns, budget estimation, shooting schedules, crew planning, visual references, and smarter location recommendations—making it a complete AI co-pilot for film production.
Log in or sign up for Devpost to join the conversation.