🚀 Forky – AI-Powered Open-Source Collaboration
💡 Inspiration
Forky was born during the RevolutionUC Hackathon, supported by Google Cloud's #VertexAISprint program. Our inspiration came from a common challenge in open source: the overwhelming complexity of joining new projects.
Many developers want to contribute but get lost in:
- Large, complex codebases that are hard to navigate.
- Finding good first issues that match their skill level.
- Understanding contribution guidelines and best practices.
We wanted to make open-source contribution more approachable and fun by leveraging AI to:
✅ Help developers understand codebases quickly
✅ Identify beginner-friendly issues
✅ Provide intelligent suggestions for contributions
✅ Make the entire experience more interactive and less intimidating
🛠️ How We Built Forky
Forky is powered by a combination of cutting-edge AI techniques designed to streamline open-source collaboration:
🔹 RAG (Retrieval-Augmented Generation) – Enhances Forky’s ability to analyze repositories and extract the most relevant information for contributors.
🔹 AI Agents – Provide real-time PR insights, smart issue recommendations, and repository analysis.
🔹 GitHub API Integration – Fetches repository data, including issues, pull requests, and contributor activity.
🔹 Gemini 2.0 – Enables intelligent code explanations and natural-language interaction with repositories.
🔹 Visualization Tools – Generates interactive diagrams to help contributors understand code structure.
🔹 Flask & Jinja2 – Backend framework for handling AI requests and chat interactions.
🚀 Key Features
📌 Smart Repository Analysis
🔹 Uses RAG (Retrieval-Augmented Generation) to understand repository structure.
🔹 Generates comprehensive project summaries for new contributors.
🔹 Creates visual diagrams of code architecture to simplify repo navigation.
📌 Issue Classification & Recommendations
🔹 Automatically categorizes issues by difficulty (Beginner, Intermediate, Advanced).
🔹 Suggests beginner-friendly starting points for new contributors.
🔹 Generates creative contribution ideas to help developers get involved.
📌 Interactive AI Chat
🔹 Real-time code explanations – Understand functions, dependencies, and logic instantly.
🔹 Implementation guidance – Get AI-driven suggestions for improving pull requests.
🔹 Instant PR insights – Forky reviews PRs and suggests improvements before submission.
🎯 Challenges We Faced
1️⃣ Handling large repositories efficiently – We optimized retrieval methods to extract key insights quickly.
2️⃣ Balancing AI accuracy and usability – We fine-tuned Forky's responses to provide precise yet concise explanations.
3️⃣ Creating meaningful issue classifications – AI needed proper training to distinguish between "easy" and "hard" issues.
4️⃣ Ensuring real-time interactivity – We improved response times to make Forky feel more conversational and engaging.
🚀 What We Learned
Throughout this journey, we gained valuable insights into:
✅ How AI can improve open-source onboarding and code comprehension.
✅ The importance of user experience in AI-assisted development tools.
✅ Optimizing LLMs for structured repository analysis and interactive feedback.
Most importantly, we learned that open-source contribution should be easy, accessible, and fun! 🎉
📍 Next Steps
🔹 Expand Forky’s capabilities with AI-powered code refactoring suggestions.
🔹 Improve issue-matching algorithms to help contributors find the perfect tasks.
🔹 Build a browser extension for inline GitHub assistance.
🔹 Launch Forky as a GitHub app to integrate directly with developer workflows.
🚀 Forky is here to make open-source collaboration smarter, faster, and more accessible. Join us in shaping the future of open-source development!
Log in or sign up for Devpost to join the conversation.