Back to projects

DevScope

DevScope is an AI-powered app that analyzes a GitHub user's profile and repositories, then returns mentorship-style feedback on their portfolio.

HTML/CSS/JSFastAPILangGraphGroq
Visit LiveSource Code
Mentor-style feedback
GitHub API enrichment
FastAPI + Vanilla JS split
Render deployment
DevScope preview

Project Summary

DevScope analyzes a GitHub profile and repository set, then converts that information into practical, mentorship-style feedback.

The app is designed to help students understand how their public work looks to a reviewer and what they can improve next.

What You Used

This stack intentionally keeps the UI simple while letting the AI pipeline do the heavy lifting.

  • Frontend: Vanilla HTML, CSS, and JavaScript.
  • Backend/API: FastAPI (Python) with Uvicorn.
  • AI / Orchestration: LangGraph and LangChain.
  • LLM: Groq running llama-3.1-8b-instant.
  • Data and API integration: GitHub REST API endpoints for user and repository data.
  • Other libraries: requests and python-dotenv.

Hosting / Deployment

The app is deployed as two separate services so the UI and API can scale and fail independently.

  • Platform: Render.
  • Deployment model: one backend API service and one static frontend service.
  • Configuration: render.yaml with GROQ_API_KEY, GITHUB_TOKEN, and BACKEND_URL environment variables.

Core Features

The product is focused on fast feedback and a clear review experience.

  • Enter a GitHub username and analyze the portfolio.
  • Fetch recent repositories, languages, and public repository count.
  • Generate AI feedback with actionable suggestions.
  • Handle API and LLM issues with timeouts, rate limits, and retries.
  • Present a clean UI with metrics and expandable review sections.

Architecture

The high-level flow is simple: collect, enrich, review, and render.

  • The user submits a GitHub username in the frontend UI.
  • The UI calls a FastAPI POST /review endpoint.
  • A LangGraph pipeline runs extract_github_data against the GitHub API, then code_mentor_review through the Groq LLM.
  • The backend returns JSON containing extracted data and mentor feedback.
  • The UI renders metrics and the mentor review response.

Portfolio-ready Blurb

Built a full-stack AI portfolio reviewer using Vanilla JS, FastAPI, LangGraph, and Groq Llama 3.1. The app consumes the GitHub REST API to extract repository and language insights, then generates mentorship-style feedback through an LLM pipeline. Deployed as separate frontend and backend services on Render, with production-style handling for rate limits, timeouts, and API errors.

"Built a full-stack AI portfolio reviewer using Vanilla JS, FastAPI, LangGraph, and Groq Llama 3.1. The app consumes the GitHub REST API to extract repository and language insights, then generates mentorship-style feedback through an LLM pipeline. Deployed as separate frontend and backend services on Render, with production-style handling for rate limits, timeouts, and API errors."