Case study / AI-powered developer tool
EggScan
A developer tool combining GitHub data, language models, and human-readable output.
Overview
What the system is
EggScan instantly scans, audits, and analyzes any GitHub profile with Groq-powered AI, grading it on a 0–100 Egg Score and returning both a humorous roast and professional, constructive feedback.
Problem
The constraint that shaped it
Turn a raw GitHub profile into a scored, readable audit with both a roast and constructive feedback.
Capabilities
What it does
- 01GitHub profile extractionFetches bio, pinned repositories, language distribution, and real-time contribution statistics.
- 02GraphQL-optimized queriesRetrieves deep metrics in a single network round-trip via GitHub's GraphQL API.
- 03Groq-powered AI auditRuns LLMs on Groq's high-speed inference engine to analyze portfolio strength and coding patterns.
- 04Egg Verdict systemFive egg-themed verdicts from Golden Egg (80–100) down to Scrambled (0–24).
- 05Technical verdictCombines a light-hearted roast with actionable feedback on quality and presentation.
- 06Glassmorphic dashboardDark-themed Vite + Tailwind UI with interactive feedback cards and live loaders.
Engineering
Decisions on record
- Architecture
- Monorepo with a decoupled Spring Boot 3.3.4 (Java 21) backend and a Vite + React frontend.
- Endpoints
- /api/scan for analysis and /api/health for readiness checks.
- Configuration
- GITHUB_TOKEN bypasses rate limits; GROQ_API_KEY drives fast LLM queries.