A powerful document ingestion and Retrieval Augmented Generation (RAG) platform for intelligent document interactions

• 🔄 Workflow Management: Create and manage custom workflows with prompt templates and model configurations
• 📄 Document Processing: Intelligent document ingestion with chunking and embedding
• 💬 Contextual Chat: Chat with your documents using advanced RAG technology
• 🗄️ Vector Storage: Powered by Chroma vector database for efficient similarity search
• 👥 Session Management: Organize conversations with persistent session storage
• 🔌 API-First: RESTful API architecture for seamless integrations

graph LR
    A[Frontend UI<br/>Next.js + React] <--> B[FastAPI Backend<br/>Python]
    B <--> C[Chroma Vector DB<br/>Embeddings]
    B <--> D[Supabase<br/>Auth + Database]
    B <--> E[Gemini API<br/>LLM + Embeddings]

Before getting started, ensure you have:

• Python 3.9+ installed
• Node.js 18+ and npm/yarn
• Supabase account with API keys
• Google Gemini API key
• Docker (optional, for containerized deployment)

1. Clone the repository

   git clone https://github.com/Yash-pede/aiplanet.git
   cd aiplanet

2. Backend Setup

   cd backend
   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   pip install -r requirements.txt

3. Environment Configuration

   cp .env.example .env

   Update your .env file:

   SUPABASE_URL=your_supabase_url
   SUPABASE_ANON_KEY=your_supabase_anon_key
   SUPABASE_SERVICE_KEY=your_supabase_service_key
   GOOGLE_API_KEY=your_gemini_api_key
   CHROMA_PATH=./chroma_db
   TEMP_DIR=./temp
   DEBUG=true

4. Start the Backend

   uvicorn app.main:app --reload --host 0.0.0.0 --port 8000

5. Frontend Setup

   cd ../client
   npm install
   npm run dev

1. Navigate to the workflows section
2. Define your prompt template
3. Select your preferred LLM and embedding models
4. Configure document sources
5. Save and activate your workflow

1. Upload documents through the UI or API
2. Documents are automatically:
   • Downloaded and parsed
   • Split into meaningful chunks
   • Embedded using Gemini
   • Stored in Chroma vector database

1. Create a new chat session
2. Select your workflow
3. Ask questions about your documents
4. Get contextual responses powered by RAG

For detailed API documentation, visit /docs when running the backend server.

aiplanet/
├── backend/                 # FastAPI backend
│   ├── app/
│   │   ├── api/            # API routes
│   │   ├── services/       # Business logic
│   │   ├── dao/           # Data access objects
│   │   ├── clients/       # External service clients
│   │   ├── schemas/       # Pydantic models
│   │   └── core/          # Configuration and utilities
│   ├── requirements.txt
│   └── .env.example
├── client/                 # Next.js frontend
│   ├── app/               # App router pages
│   ├── components/        # React components
│   ├── hooks/            # Custom React hooks
│   ├── lib/              # Utility libraries
│   └── utils/            # Helper functions
├── docker-compose.yml     # Docker configuration
└── README.md

Key configuration options in app/core/config.py:

• APP_NAME: Application name
• API_PREFIX: API route prefix
• CORS_ORIGINS: Allowed CORS origins
• DEBUG: Debug mode toggle

graph TD
    A[Document Upload] --> B[Metadata Storage]
    B --> C[File Download]
    C --> D[Document Parsing]
    D --> E[Text Chunking]
    E --> F[Generate Embeddings]
    F --> G[Store in Chroma]
    G --> H[Mark as Processed]

graph TD
    A[User Message] --> B[Similarity Search]
    B --> C[Retrieve Context]
    C --> D[Construct Prompt]
    D --> E[LLM Generation]
    E --> F[Save Response]
    F --> G[Return to User]

• Authentication: Supabase-based user authentication with JWT tokens
• Authorization: Role-based access control per user
• CORS: Configurable cross-origin resource sharing
• Input Validation: Comprehensive request validation using Pydantic
• Error Handling: Secure error responses without sensitive data exposure

• Support for additional file formats (Word, HTML, Images)
• Dynamic model switching during conversations
• Enhanced UI for workflow visualization
• Improved error handling and user feedback

• Advanced admin dashboard
• Parallel document processing
• Caching layer for improved performance
• Webhook support for integrations

• Multi-tenant architecture
• Custom model fine-tuning
• Advanced analytics and insights
• Enterprise SSO integration