• 

  image 1

• 

  image 2

• 

  image 3

• 

  image 4

• 

  image 5

Inspiration

In the film industry, production research is a massive bottleneck. Major studios have entire departments dedicated to verifying locations, historical accuracy, and logistical costs. Independent filmmakers have to do this themselves across dozens of browser tabs, and making a mistake can derail a production. I wanted to turn this manual, tedious process into an autonomous intelligence workflow.

What it does

CineScout is an autonomous research platform that helps filmmakers move from idea to greenlight with evidence. When a user submits a production query, CineScout dispatches concurrent AI research workers to pull real-time "ground truth" from the live web. Instead of just generating text, it provides verifiable data, sources, and intelligence to back up production decisions.

How we built it

I built the architecture combining three core technologies:

• Google Gemini: Serves as the primary reasoning engine, planning the research strategy and synthesizing the final reports.
• Google Cloud Agent Development Kit: Handles the pipeline orchestration, managing the lifecycle and handoffs between different agentic workers.
• Parallel API: Powers the live web intelligence and provenance, ensuring the data the agents retrieve is current, accurate, and properly sourced (which is critical for the Parallel Search Track).

Challenges we ran into

One of the biggest hurdles was orchestrating multiple agents without them getting stuck in loops or hallucinating facts. Getting the Google Cloud Agent Development Kit to seamlessly hand off the planning phase to the Parallel Search execution phase required careful prompt engineering and precise data validation.

Accomplishments that we're proud of

I am incredibly proud of achieving true parallel execution. Seeing the platform dispatch multiple search queries simultaneously and synthesize the live results into a single, cohesive production report using Gemini is magical. It genuinely feels like having a team of research assistants working in the background.

What we learned

I deepened my understanding of agentic orchestration. Building a single-prompt LLM wrapper is easy, but building a multi-agent system where agents verify each other's work using live web data taught me a lot about building robust, production-ready AI applications.

What's next for CineScout

The immediate next step is integrating a budgeting agent that can automatically generate cost estimates based on the Parallel search data. I also plan to add a script-breakdown feature where filmmakers can upload a PDF, and CineScout will automatically begin researching the locations and props mentioned in the scene and execution. Seeing the platform dispatch multiple search queries simultaneously and synthesize the live results into a single, cohesive production report using Gemini is magical. It genuinely feels like having a team of research assistants working in the background.

Built With

• agent-development-kit
• github
• google-cloud
• google-gemini
• parallel-search-api
• python
• render