Librai - AI-Powered Document Processing & Chat Platform

Overview

Librai is a monorepo containing a full-stack application that enables users to process documents and interact with them through an AI chatbot interface. The platform consists of two main applications:

• librai-server: A document processing service that handles file uploads, text extraction, and vector embeddings
• librai-ui: A modern chat interface for interacting with processed documents using AI

Key Features

Document Processing (librai-server)

• File upload support with progress tracking
• Support for multiple file formats:
  • PDF (with page range exclusion)
  • EPUB (with section exclusion)
  • TXT
  • MD
• Smart text chunking with metadata preservation
• OpenAI embeddings generation (text-embedding-3-large)
• Qdrant vector database integration
• Express.js REST API

Chat Interface (librai-ui)

• Interactive AI chatbot with context-aware responses
• Multiple theme support:
  • Charcoal (light/dark)
  • Leaf (light/dark)
  • Twilight (light/dark)
  • Volcano (light/dark)
  • Ocean (light/dark)
• UI built with SvelteKit
• User authentication via PocketBase
• Conversation management:
  • Save and resume conversations
  • Delete conversations
  • Share via public links
  • Edit conversation titles
• Rate limiting with Upstash Redis
• Real-time streaming responses
• Markdown and LaTeX support
• Responsive design with Tailwind CSS

Getting Started

1. Clone the repository:

git clone https://github.com/yourusername/librai.git
cd librai

2. Install dependencies:

pnpm install

3. Set up environment variables: Create .env files in both apps/server and apps/ui directories using their respective .env.example files as templates.

4. Start the development servers:

pnpm dev

Environment Variables

Server (.env)

PORT=3000
OPENAI_API_KEY=your_api_key
OPENAI_EMBEDDINGS_MODEL=text-embedding-3-large
QDRANT_API_URL=your_qdrant_url
QDRANT_API_KEY=your_qdrant_key
QDRANT_COLLECTION=your_collection_name

UI (.env)

# OpenAI Configuration
PRIVATE_CHAT_API_KEY=your_openai_key
PRIVATE_EMBEDDINGS_API_KEY=your_openai_key
PRIVATE_EMBEDDINGS_MODEL=text-embedding-3-large
PRIVATE_CHAT_MODEL=gpt-4o-mini
PRIVATE_CHAT_BASE_URL=your_openai_base_url

# Qdrant Configuration
PRIVATE_QDRANT_ENDPOINT_URL=your_qdrant_url
PRIVATE_QDRANT_COLLECTION_NAME=your_collection_name
PRIVATE_QDRANT_API_KEY=your_qdrant_key

# System Configuration
PRIVATE_SYSTEM_PROMPT=your_custom_prompt

# Redis Configuration
PRIVATE_UPSTASH_REDIS_REST_URL=your_upstash_url
PRIVATE_UPSTASH_REDIS_REST_TOKEN=your_upstash_token

# PocketBase Configuration
VITE_POCKETBASE_URL=http://127.0.0.1:8090

# Application Settings
PUBLIC_APP_TITLE=Librai
PUBLIC_APP_DESCRIPTION=Your app description
PUBLIC_APP_SHORT_TITLE=Librai
PUBLIC_THEME=charcoal
PUBLIC_APP_URL=your_app_url

# Chatbot Configuration
PUBLIC_CHATBOT_DESCRIPTION=Your chatbot description
PUBLIC_CHATBOT_THINKING_TEXT=Your thinking text

# OpenGraph/Social Media
PUBLIC_APP_OG_IMAGE=your_og_image_url
PUBLIC_APP_TWITTER_IMAGE=your_twitter_image_url

Documentation

• Server Documentation
• UI Documentation

License

This project is licensed under the GNU Affero General Public License v3.0 (AGPL-3.0).