This project provides a free, one-click deployable private Gemini chatbot application, targeting users who want to leverage Google's Gemini models (1.5 Pro, 1.5 Flash, Pro, Pro Vision) with a user-friendly interface. It offers a web application, a cross-platform desktop client, and supports multimodal capabilities, plugins, and extensive Markdown rendering.

How It Works

The application is built with Next.js, Tailwind CSS, and shadcn/ui, providing a responsive and feature-rich user experience. It leverages the Gemini API for natural language processing and supports multimodal inputs like images and videos. Function calling is integrated for plugin support, enabling features like web search, reading, and more. Data is stored locally in the browser for privacy.

Quick Start & Requirements

• Web Deployment: One-click deployment via Vercel or Cloudflare.
• Desktop App: Downloadable clients for Windows, macOS, and Linux from releases.
• Docker: docker pull xiangfa/talk-with-gemini:latest and docker run -d --name talk-with-gemini -p 5481:3000 xiangfa/talk-with-gemini.
• Prerequisites: NodeJS >= 18, Docker >= 20.
• API Key: A Gemini API key is required.
• Docs: https://github.com/u14app/gemini-next-chat

Highlighted Details

• Supports Gemini 1.5 and Gemini 2.0 multimodal models, including image and video understanding.
• Includes built-in plugins for web search, web reading, Arxiv search, and weather.
• Offers a cross-platform desktop client that can reside in the menu bar for efficiency.
• Features full Markdown support, including KaTex, code highlighting, and Mermaid charts.

Maintenance & Community

The project is actively maintained, with recent releases and ongoing development indicated by roadmap items. Community contributions are welcomed via pull requests and issue reporting.

Licensing & Compatibility

Licensed under the MIT License, allowing for commercial use and integration with closed-source projects.

Limitations & Caveats

The Multimodal Live API currently only supports the Gemini 2.0 Flash model and may require a Cloudflare Worker proxy for access in certain regions (e.g., China). Chinese voice output is not yet supported for the Multimodal Live API.