shamspias / langchain-gemini-api

An innovative AI conversation API leveraging Google's Gemini for multimodal understanding. Combines FastAPI, Langchain, and Redis for robust, scalable, and privacy-conscious text and image-based interactions

langchain-gemini-api

Overview

langchain-gemini-api is an AI-powered conversation API that integrates Google's Gemini API, designed to facilitate advanced text and image-based interactions. This project combines the capabilities of modern deep learning models with FastAPI for high performance and scalability, Langchain for sophisticated conversational workflows, and Redis for ephemeral conversation storage.

Key Features

• Asynchronous API: Utilizing FastAPI for enhanced performance and scalability.
• Langchain Integration: Advanced conversational workflows using multiple AI models.
• Google Gemini API: Incorporating Gemini, Gemini Pro, and Gemini Pro Vision for superior conversation understanding and generation.
• Ephemeral Conversation Storage: Implementing Redis for data privacy and efficient memory usage.
• User-Friendly Interface: Simplified API endpoints for easy integration.

Getting Started

Prerequisites for Google Gemini API

• Get a Google Gemini AI API key here and add it to Secrets in your Repl with the key GOOGLE_API_KEY.

Prerequisites for langchain-gemini-api

• Python 3.9 or higher
• FastAPI
• Uvicorn
• Redis
• Access to Google Gemini API

Installation

Install Redis

1. Install Redis:

   sudo apt update
   sudo apt install redis-server

   Windows users can download Redis from here
2. Test Redis:

   redis-cli ping

   If Redis is running, it will return PONG.

Install Python Dependencies

1. Clone the repository:

   git clone https://github.com/shamspias/langchain-gemini-api.git

2. Create a virtual environment:

   python3 -m venv venv

   for Windows:

   python -m venv venv

3. Activate the virtual environment:

    source venv/bin/activate

4. Navigate to the project directory and install dependencies:

   cd langchain-gemini-api
   pip install -r requirements.txt

Configuration

1. Create a .env file in the project directory and add the following:

   GEMINI_API_KEY=<your_google_gemini_api_key> # e.g. "1234567890"
   REDIS_URL=redis://localhost:6379/0 # default Redis URL
   MY_API_KEY=<your_api_key_to_access_your_backend_api> # e.g. "1234567890"
   SYSTEM_INSTRUCTION=<your_system_instruction_for_gemini> # e.g. "Your are Sam and an AI assistant...etc"

Running the API

1. Start fastapi server:

uvicorn app.main:app --reload

The API will be available at http://localhost:8000.

API Endpoints

• /conversations/<conversation_id>: Endpoint for text and image based conversations.
• /delete/<conversation_id>: Endpoint to delete conversation data.

API Documentation

• /docs: Swagger UI for API documentation.
• /redoc: ReDoc UI for API documentation.

Acknowledgments

• Google Gemini API team
• Langchain contributors
• FastAPI community

Demo

Conversation

With telegram bot use this one: https://t.me/rim_nova_bot

This is the telegram bot repo: Gemini-Telegram-Bot

Video

• Coming soon