Thread AI - Backend

This is the backend service for Thread AI - Smart Email Reply Generator.
It acts as a unified LLM router to handle multiple AI providers (OpenAI, Gemini, and Ollama) and generate context-aware, high-quality email replies.

Features

• Multi-LLM Support – OpenAI, Gemini, and Ollama in one place.
• Dynamic Provider Selection – Switch AI providers via API.
• Prompt Engineering – Centralized and optimized prompts for better responses.
• FastAPI Backend – Lightweight and high-performance.
• Easy Deployment – Works locally or on cloud.

Project Structure

├── llm_router.py # Routes requests to the correct LLM provider
├── main.py # FastAPI app entry point
├── prompt.py # Contains the system & user prompt templates
├── providers/
│ ├── gemini_provider.py # Google Gemini integration
│ ├── ollama_provider.py # Ollama local LLM integration
│ └── openai_provider.py # OpenAI API integration
├── requirements.txt # Python dependencies

Installation

1 Clone the Repository

git clone https://github.com/ashankgupta/threadwise-ai-backend.git
cd thread-ai-backend

2 Create Virtual Environment

python -m venv venv
source venv/bin/activate   # For Linux/Mac
venv\Scripts\activate      # For Windows

3 Install Dependencies

pip install -r requirements.txt

Running the Server

uvicorn main:app --reload

Server will start at:

http://127.0.0.1:8000

API Endpoints

Method	Endpoint	Description
POST	/generate	Generate an AI email reply
GET	    /providers	List available AI providers
POST	/switch	    Switch active AI provider

License

This project is licensed under the MIT License.