Skip to content
Jacob Clifton
← All work

Independent · 2024 to 2025

Koobrik

A serverless tool that reads a comic book page by page and synthesizes it into studio-style coverage.

Role · Solo. Architecture and build.

The problem

Comics and graphic novels are a significant source of adapted IP, but evaluating them is slow. The story lives in the interaction between art and text, so neither a transcript nor a script summary tells you what you need to know.

Approach

  1. 01

    Extracts both text and visual description from each PDF page, so the page is represented as it reads rather than as it transcribes.

  2. 02

    Generates a structured per-page JSON record covering dialogue, setting, characters, and themes.

  3. 03

    Synthesizes the page-level record into cohesive coverage, so every claim in the summary traces back to a specific page.

Outcome

Working serverless pipeline that turns a full-length comic PDF into structured, page-traceable coverage.

Stack

PythonGoogle Vertex AIAWS LambdaPyPDF2Pandas