Chatsource

This project offers APIs for managing chatbots with Retrieval-Augmented Generation (RAG), using OpenAI’s LLMs and embeddings. Users can create, train, and query chatbots, with support for document-based training and multi-user management.

Features

• User Authentication: User-based authentication with support for managing multiple chatbots per user.
• Chatbot Creation: Users can create chatbots by specifying their parameters such as language model, temperature, and instructions.
• Document-Based Training: Chatbots can be trained with documents provided by the user. Each document is linked to a chatbot and is used for enhancing its capabilities.
• Custom Query Engine: Users can query their chatbots, and the system will generate responses using OpenAI's models based on the trained documents.
• Embedding Support: The system uses OpenAI's embedding model for processing documents and queries.
• Vector Indexing: A vector store is used to enhance the query engine's efficiency in generating responses.

Tech Stack

• FastAPI: Fast and modern web framework for building APIs with Python.
• OpenAI API: Utilizes OpenAI's models for both generating responses and embedding content.
• LlamaIndex: Used to create and manage a vector store index for efficient document querying and retrieval.

Installation

1. Clone the repository:

    git clone https://github.com/tunahsu/chatsource-api.git
    cd chatsource-api

2. Install dependencies:

   pip install -r requirements.txt

3. Set up environment variables: Create a .env file in the project root and add the necessary environment variables (e.g., OpenAI API key, database connection).

4. Run the development server:

   python main.py