SUNSETSEJONG UNIVERSITY
← Back to projects
P-08 · Supporting

roastivator

SupportingWorking PrototypeWeb
roastivator Next.js web application accepting a GitHub username, fetching public profile and repository signals, applying deterministic rules on commits, repository names, activity, stars, forks, followers, and emojis, and rendering humorous roasting output through a step-based flow.SIG-RST · PROFILE ROASTINPUTUsernamestep formGITHUB SIGNALSCommitsmessagesReposnames · starsActivityforks · emojisFollowerspublic APIRULE ENGINERule setsdeterministicRandom pickvariationOUTPUTLoading stepmotion UIRoast resultplayful copyrules-based · not model-backed AI
System diagram

Verified summary

Web application that analyzes public GitHub profile signals and generates humorous roasting output through rules.

Problem or experiment question

Can commit messages, repository names, and activity patterns produce entertaining feedback from public GitHub data?

What was actually built

  • Next.js, React, and TypeScript client
  • Step-based input, loading, and result flow
  • Rules based on commit messages, repository names, stars, forks, followers, and emojis
  • Error-boundary and hook structure with Jest configuration

Important technical decisions

  • Rules-based engine keeps output inspectable and reproducible
  • Step flow separates loading state from result presentation
  • Public GitHub API signals chosen as the only data source

Architecture or workflow

  • Input step collects a public GitHub username
  • Analysis step applies deterministic rules to profile signals
  • Result step presents generated roasting output

Known limitations

  • Core output is deterministic and rule-based, not model-backed AI
  • Random selection may make repeated results inconsistent
  • Public GitHub API rate limits may affect usage

Repositories

Related projects