Web app for Gemini multimodal chat, deployable in seconds
Top 44.3% on sourcepulse
This project provides a simple, deployable web interface for interacting with Google's Gemini multimodal models, targeting users who want a quick and accessible way to leverage Gemini's capabilities without complex setup. It offers a user-friendly chat experience, including microphone, camera, and screen sharing features, and acts as an API proxy to the OpenAI format, enabling integration with various AI clients.
How It Works
The project utilizes Deno and Cloudflare Workers for a serverless, free deployment. It acts as a proxy, converting Gemini's API requests and responses into the more universally compatible OpenAI format. This approach allows for direct access from various regions and networks, bypassing potential geographical restrictions, and facilitates integration with tools like ChatBox and Cursor.
Quick Start & Requirements
Highlighted Details
Maintenance & Community
The project is maintained by "技术爬爬虾" (Tech Climbing Shrimp), with contributions and inspirations from ChrisKyle and PublicAffairs. Links to Bilibili, YouTube, Douyin, and WeChat official accounts are provided for community engagement.
Licensing & Compatibility
The project's UI and Gemini-to-OpenAI conversion components are licensed under the MIT License. This license permits commercial use and linking with closed-source projects.
Limitations & Caveats
Users in certain regions (e.g., Guangdong/Hong Kong) using Cloudflare Workers might encounter "400: User location is not supported" errors due to CDN routing; Deno deployment is recommended to mitigate this.
1 month ago
Inactive