🧠 AnyBrain

AnyBrain is an open-source, local RAG (Retrieval-Augmented Generation) tool designed to effortlessly create and query embeddings from your personal documents. Turn any folder of files into an intelligent, queryable brain using the power of LLMs.

✨ Features

• Universal Ingestion: Automatically processes multiple formats:
  • 📄 PDF (Digital & Scanned via GPU-accelerated OCR)
  • 📝 Word (.docx) & Text (.txt, .md)
  • 📊 Excel (.xlsx)
  • 🖼️ Images (.jpg, .png) containing text.
• Model Agnostic: Seamlessly switch between Ollama (Local), Google Gemini, DeepSeek, or OpenAI.
• Privacy First: Vector processing and storage happen locally on your machine using ChromaDB.
• Modern Interface: Built with Chainlit, featuring Neural Text-to-Speech (TTS) and dynamic settings.

🚀 Quick Start

No complex setup required. AnyBrain handles the environment for you.

1. Clone the repository (or download as ZIP).
2. Add your files to the data/inputs folder.
3. Double-click start.bat.

The script will automatically:

• Check for Python.
• Create a virtual environment (venv).
• Install all dependencies (CUDA enabled).
• Ingest your documents.
• Launch the Chat Interface in your browser.

🛠️ Configuration

Once the interface is running, click on the Settings (⚙️) icon to:

• Switch AI Providers (Local vs Cloud).
• Enter API Keys (stored securely in session).
• Change the assistant's personality.

📦 Requirements

• OS: Windows (Batch script provided).
• Python: 3.10 or higher.
• GPU (Optional): NVIDIA GPU recommended for faster OCR and Embedding generation.

📂 Project Structure

AnyBrain/
├── data/
│   ├── inputs/          # Drop your files here
│   ├── vector_db/       # Local Knowledge Base (Chroma)
│   └── history/         # Saved chat sessions
├── src/                 # Source code
├── start.bat            # Auto-installer & Launcher
└── requirements.txt     # Python dependencies