Sufian

ReproForge AI — From Failing Tests to Verified Fixes ReproForge AI is an AI-powered debugging assistant that helps developers identify software bugs, generate code repairs, and verify their solutions through automated testing. As a computer science student, I've experienced how frustrating debugging can be, especially when AI-generated solutions appear correct but haven't actually been tested. I built ReproForge AI to address that problem. The application follows a three-step workflow: 1. Reproduce:Run automated Python tests using pytest to identify failing functionality. 2. Repair: Use Meta Muse Spark 1.3 through the Meta Model API to analyze the broken code and generate a proposed fix. 3. Verify: Run the same tests against the repaired code and display the results alongside the code changes. In a live demonstration, ReproForge corrected a broken discount calculation, improving the results from 3 failing tests to 4 passing tests. Built using Python, Streamlit, pytest, and Meta Muse Spark, ReproForge is an early-stage prototype demonstrating how generative AI and automated software testing can work together to support more reliable debugging.