Reconstruct Codebase Architecture Instantly.
RepoMind AI is an AI-powered developer workspace that helps developers understand repository architecture in seconds. By parsing remote directory structures and configuration files recursively, it reconstructs modular flowcharts and folder hierarchies without requiring local repository cloning, Docker containers, or resource-intensive vector indexing.
| Feature | ✕ Vector Indexing (Traditional) | ✓ Metadata Mapping (RepoMind) |
|---|---|---|
| Repository Access | Typically requires local cloning and indexing | No local cloning required (uses GitHub APIs) |
| Setup Time | Minutes/Hours (embedding models, database setup) | Instant (queries public trees on demand) |
| Token Cost | Extremely High (feeds entire file chunks into LLM) | Extremely Low (sends only topology trees & manifests) |
| Security Risk | High (stores raw source code in vector indexes) | Minimal (reads only manifests; stores keys in browser storage) |
- 🗺️ Interactive Flow Canvas: Renders fully zoomable and pannable AI-generated Mermaid.js architecture diagrams. Export visuals directly as vectors (SVG) or high-res images (PNG), copy the raw Mermaid code, or view in solid, opaque Fullscreen.
- 📁 Topology Tree & Code Inspector: Explore a nested directory structure filtering out binaries, libraries, and caches. Click any file to slide open the Source File Inspector drawer and inspect code files directly using GitHub APIs.
- 📦 Manifest Inspector: A side-by-side split screen viewer that displays package configuration files (such as
package.jsonorrequirements.txt) with line numbers. - 📊 Ecosystem Composition Dashboard: A visual dashboard showcasing codebase maintainability scorecards, language distribution bars, noise reduction ratios, and real-time execution statistics.
- 🔗 Shareable Permalinks: Share scans directly using permalink parameters (
?repo=<owner>/<repo>&branch=<name>). Landing on the workspace with parameters automatically triggers codebase scanning. - 🔒 API First & Secure: All credentials (Gemini keys, GitHub tokens) are stored locally in the browser's
localStorageand never saved on the server. Mermaid XSS protection is set tostrict.
- Frontend: Next.js 16 (App Router), Tailwind CSS v3, TypeScript, Mermaid.js, Lucide Icons
- Backend: FastAPI, Uvicorn, Google GenAI SDK (Python 3.10+)
- Node.js (v18+)
- Python 3.10+
Navigate to the backend/ directory:
cd backendCreate a virtual environment and activate it:
- Windows (PowerShell):
python -m venv .venv .\.venv\Scripts\Activate.ps1 - macOS / Linux:
python3 -m venv .venv source .venv/bin/activate
Install dependencies:
pip install -r requirements.txtCreate your .env configuration:
cp .env.example .env(Optionally paste your GEMINI_API_KEY and GITHUB_TOKEN inside .env to load them automatically, or pass them inside the Web UI settings).
Launch the backend server:
python main.pyThe server will start running on http://127.0.0.1:8000.
Return to the root workspace directory and install Node packages:
npm installLaunch the development workspace:
npm run devOpen http://localhost:3000 in your browser.
RepoMind is configured for immediate cloud deployment using direct native git integrations:
- Deploy the
/backendfolder. - Select Python 3 runtime.
- Build Command:
pip install -r requirements.txt - Start Command:
python main.py - Configure the following Environment Variables in the Render Dashboard:
ENVIRONMENT=productionGEMINI_API_KEY=your_gemini_api_key(Google AI Studio Key)GITHUB_TOKEN=your_github_token(Optional GitHub PAT)
- Import your repository into Vercel (Preset: Next.js).
- Configure the following Environment Variable in the Vercel Dashboard:
NEXT_PUBLIC_API_URL=https://your-backend-live-url.onrender.com(Your live Render backend address)
- Deploy! Vercel will optimize and serve your frontend pages statically.