Gitsby Project Management Platform

Gitsby is a platform designed to help teams track GitHub commits, collaborate effectively, and leverage AI-driven insights. By integrating with GitHub, utilizing tRPC for API routes, and providing powerful tools for meeting transcriptions, Q&A, and team management, Gitsby brings all project elements into one seamless experience.

Try out it out from here - https://gitsby-psi.vercel.app/

Features

• Dashboard: A central hub displaying a list of projects, recent commits, AI-generated summaries, and project details.
• Q&A Section: Users can ask project-related questions, with AI-generated responses based on the project’s codebase and summaries.
• Meetings: Upload meeting recordings to transcribe and generate summaries via AssemblyAI.
• Project Creation: Create new projects by linking to GitHub repositories, automatically importing commits, and generating AI-driven summaries and code embeddings for querying.
• Member Management: Invite and manage project members, assigning roles for better team collaboration.
• Archive Projects: Archive completed projects to keep them accessible for future reference while maintaining a streamlined active project list.

Tech Stack

Frontend:

• Next.js
• Tailwind CSS
• ShadCN

Backend:

• tRPC (for API routes)

Database:

• PostgreSQL with Prisma ORM for handling database interactions.
• pgvector for storing vector embeddings of codebase summaries and commits for efficient querying.

External Services:

• Langchain (for extracting files from GitHub repositories)
• GitHub API (via Octokit)
• Gemini API (for generating AI summaries)
• AssemblyAI (for meeting transcription)

Process Flow Summary

Here’s how the platform works step-by-step:

1. Sign-in → Users authenticate via Clerk.
2. Dashboard → Displays all active projects with recent commits and AI-generated summaries.
3. Create Project → Users input a GitHub URL to create a project. Prisma interacts with the database to store project details, and Langchain extracts repository files.
4. Commits Processing → GitHub commits are fetched and stored in the database. Gemini API generates summaries, which are then stored in pgvector for efficient querying.
5. Q&A → Users can ask questions related to the project’s codebase, and tRPC routes fetch AI-generated answers based on the stored summaries and codebase.
6. Meeting Transcription → Users upload meeting audio files, which AssemblyAI transcribes and summarizes. The transcriptions are stored in the database.
7. Member Management → Users invite team members and manage roles. Prisma is used to update member data in the database.
8. Archive Project → Completed projects are archived but remain accessible for future reference in the database.