Case Study - The AI backend that analyses screenplays for First Look

We built the backend that turns screenplay PDFs into scenes, characters and metrics, and generates professional coverage with GPT-4o.

Client
First Look
Year
Service
Backend · AI
Problem
First Look needed to turn screenplay PDFs, with very different layouts, into structured analysis and coverage that stays consistent from one script to the next.
Our role
Eduardo worked as a Software Engineer on the AI analysis backend.
What we built
  • PDF parsing into scenes, characters and script metrics
  • Coverage generated with GPT-4o on top of that structure
  • Versioned prompts, to improve results without breaking what worked
  • Calibrated scoring, so grades are comparable across scripts
  • Vision fallback when a PDF can’t be read as text
Stack
  • GPT-4o
  • MySQL
  • Linux
Result
Live at first-look.net: every script that’s uploaded gets its First-Look Score and a report with feedback.

More case studies

An adaptive English level test built for Spanish speakers

Inglexam is our own product: it estimates your English level from A1 to C2 and explains your mistakes in Spanish.

Read case study

Spanish-language trivia, generated and fact-checked by an AI pipeline

Trivias.de is our own trivia platform in Spanish: no sign-up to play, with categories, results and leaderboards.

Read case study
Contact

Have a project in mind? Let's talk.

We start with a 30-minute call: you tell us what you want to achieve, and we tell you whether we can help and how. Eduardo replies personally.