StackPilot is a full-stack AI developer assistant that enables developers to chat with GitHub repositories, technical documentation, PDFs, Markdown files, and codebases using Retrieval-Augmented Generation (RAG).
The platform combines semantic search, vector embeddings, repository indexing, documentation ingestion, and LLM-powered reasoning to provide accurate, context-aware coding assistance with verifiable source citations.
Try StackPilot at: https://stack-pilot-jet.vercel.app
- Chat with GitHub repositories using AI
- Import and index technical documentation
- Upload PDFs and Markdown files for semantic search
- Build a unified developer knowledge base
- Provide source-cited AI responses
- Reduce developer context switching
- Improve onboarding for unfamiliar codebases
- Deliver context-aware coding assistance
React + Vite + Tailwind
│
▼
Firebase Authentication
│
▼
Node.js + Express Backend
│
├── LangChain.js
├── Gemini API
├── OpenRouter API
├── GitHub REST API
├── Jina AI Reader
├── ChromaDB
└── Firestore
Located in
frontend/
- React.js
- Vite
- Tailwind CSS
- Framer Motion
- Firebase Authentication
- React Router
- Axios
Developers can ask
- Explain this project
- Explain this function
- Find where authentication is implemented
- Summarize this repository
- Explain folder structure
- Generate onboarding guide
- Explain dependencies
- Suggest improvements
Supports
- Public repositories
- README extraction
- Folder structure analysis
- Documentation indexing
- Semantic repository search
Import documentation websites including
- React
- Next.js
- Express
- Node.js
- Tailwind CSS
- Firebase
- FastAPI
- LangChain
- Python Docs
- Any developer documentation
Upload
- Technical books
- Research papers
- API documentation
- User manuals
- Notes
Search using semantic retrieval.
Users can
- Save prompts
- Edit prompts
- Delete prompts
- Organize reusable developer prompts
Store
- Previous chats
- AI responses
- Source citations
- Repository-specific conversations
Login
│
▼
Dashboard
│
▼
Import Repository
│
▼
Index Documentation
│
▼
Generate Embeddings
│
▼
Store in ChromaDB
│
▼
Chat with AI
│
▼
Receive Source-Cited Response
Located in
backend/
Responsible for
- Repository import
- README extraction
- File tree retrieval
- Repository metadata
- Code indexing
Uses
- GitHub REST API
Responsible for
- Documentation crawling
- Content extraction
- Text preprocessing
Uses
- Jina AI Reader API
Workflow
User Query
│
▼
Vector Search
│
▼
Relevant Context Retrieval
│
▼
Prompt Construction
│
▼
Gemini / OpenRouter
│
▼
Grounded AI Response
Uses ChromaDB for
- Semantic search
- Embedding storage
- Similarity search
- Context retrieval
Supports
- Gemini API
- OpenRouter Models
Handles
- Prompt engineering
- Context injection
- Response formatting
- Source citation generation
Managed using Firebase Authentication.
Supports
- Google Sign-In
- Secure token verification
- Protected API routes
User Login
│
▼
Firebase Authentication
│
▼
JWT Verification
│
▼
Express Middleware
│
▼
Protected APIs
Repository
Documentation
PDF
Markdown
│
▼
Chunking
│
▼
Embedding Generation
│
▼
ChromaDB
│
▼
Semantic Retrieval
│
▼
Gemini / OpenRouter
│
▼
Grounded AI Response
users/
repositories/
documents/
conversations/
messages/
prompts/
analytics/
settings/
StackPilot/
│
├── backend/
│ ├── controllers/
│ ├── routes/
│ ├── middleware/
│ ├── services/
│ │ ├── githubService.js
│ │ ├── ragService.js
│ │ ├── jinaService.js
│ │ ├── embeddingService.js
│ │ └── aiService.js
│ ├── firebase/
│ ├── utils/
│ ├── config/
│ ├── server.js
│ └── package.json
│
├── frontend/
│ ├── src/
│ │ ├── components/
│ │ ├── pages/
│ │ ├── hooks/
│ │ ├── context/
│ │ ├── services/
│ │ └── utils/
│
└── README.md
Displays
- Imported repositories
- Indexed documentation
- Uploaded PDFs
- Recent conversations
- Prompt library
- Storage statistics
- AI usage analytics
POST /api/v1/auth/sync
GET /api/v1/auth/profile
POST /api/v1/repositories/import
GET /api/v1/repositories
DELETE /api/v1/repositories/:id
POST /api/v1/docs/import
GET /api/v1/docs
DELETE /api/v1/docs/:id
POST /api/v1/upload
POST /api/v1/chat
GET /api/v1/dashboard
GET /api/v1/analytics
StackPilot implements
- Firebase Authentication
- Protected routes
- Input validation
- Secure environment variables
- Rate limiting
- CORS protection
PORT=
GEMINI_API_KEY=
OPENROUTER_API_KEY=
GITHUB_TOKEN=
FIREBASE_PROJECT_ID=
FIREBASE_CLIENT_EMAIL=
FIREBASE_PRIVATE_KEY=
VITE_FIREBASE_API_KEY=
VITE_FIREBASE_AUTH_DOMAIN=
VITE_FIREBASE_PROJECT_ID=
VITE_FIREBASE_STORAGE_BUCKET=
VITE_FIREBASE_APP_ID=
VITE_BACKEND_URL=
cd backend
npm install
npm run dev
cd frontend
npm install
npm run dev
Frontend
- Vercel
Backend
- Render
Authentication
- Firebase Authentication
Database
- Firebase Firestore
Vector Database
- ChromaDB
AI Models
- Gemini API
- OpenRouter API
- React.js
- Vite
- Tailwind CSS
- Framer Motion
- Node.js
- Express.js
- Firebase Firestore
- ChromaDB
- Retrieval-Augmented Generation (RAG)
- Gemini API
- OpenRouter API
- Vector Embeddings
- Semantic Search
- GitHub REST API
- Jina AI Reader API
- Firebase Authentication
- Git
- GitHub
- Postman
- Vercel
- Render
- Private GitHub repository support
- GitHub OAuth
- VS Code Extension
- Multi-repository chat
- Repository dependency graphs
- Code execution sandbox
- Team workspaces
- Model benchmarking
- AI code review
- Pull request summarization
StackPilot aims to become an AI-native developer workspace where engineers can interact with repositories, documentation, and technical resources through natural language while receiving accurate, grounded, and source-cited responses.
MIT License
StackPilot is intended as an AI-assisted developer productivity tool. AI-generated responses should be reviewed and verified before use in production environments.
StackPilot demonstrates
- Full-Stack Web Development
- Retrieval-Augmented Generation (RAG)
- Vector Database Design
- Semantic Search
- GitHub API Integration
- Documentation Indexing
- LLM Integration
- Firebase Authentication
- Modern SaaS Architecture
- Production-ready REST API Design
- AI-Powered Developer Tooling