My request to have a single course recognized was filed in 2025 and resolved in September 2026. That wait is where Tramo came from, and it's the project I submitted this week to the Hacktoberfest Open Source AI Challenge on DEV, where I published the full story.

Tramo compares the analytic program of a course a student already passed with the program of the course they want recognized, and it drafts an equivalency report for the professor who reviews the case. It runs on Gemma 4, an open weights model served locally by Ollama, so no document ever leaves the computer. The AI reads the programs and backs every topic with evidence, plain code computes the coverage and applies the 50 percent rule, and the professor can overrule any decision before signing.

The DEV post covers everything a summary can't. It tells how the first version proposed that the student retake the names of the professors, how one setting cut the analysis from five and a half minutes to one and a half, why the model never produces a number that the interface presents as a fact, and how Tramo treats every uploaded file as hostile. You'll also find the demo video, a screenshot of every screen and the commands to run it on your own machine.

I'd love for you to read it on DEV and leave your comment there, since that's where the challenge lives and where I'll be answering every question.

Read the full post on DEV

Watch the demo on YouTube

If you've ever changed careers or universities, how long did it take to get the courses you'd already passed recognized?