yeahhe365/Gemini-Nexus

Gemini Nexus 是一款面向浏览器场景的 AI 助手扩展，集成 Gemini Web、Gemini API 与 OpenAI 兼容接口，支持网页上下文、图像处理、工具调用和 MCP 浏览器控制。

What it solves

Browsers lack a native, unified interface for interacting with various AI models. Gemini Nexus bridges this gap by providing a single interface for multiple LLM providers and enabling AI to perform actual tasks within the browser environment.

How it works

The extension acts as a sophisticated interface layer that connects to various backends, including Google Gemini (via web reverse-engineering or official API), OpenAI, Anthropic, DeepSeek, and others. It uses the Chrome DevTools Protocol to allow the AI to perform agentic actions like clicking, typing, and navigating pages. It also supports the Model Context Protocol (MCP) to connect to external tools via SSE, HTTP, or WebSockets.

Who it’s for

  • Power users who want to switch between different LLM providers (like Claude, GPT, or Gemini) within a single browser interface.
  • Developers looking to experiment with agentic browser automation and MCP tool integration.
  • Researchers exploring the capabilities of browser-native AI workflows.

Highlights

  • Multi-provider support including Gemini Web, OpenAI, Anthropic, DeepSeek, and OpenRouter.
  • Agentic browser control using Chrome DevTools Protocol for navigation and interaction.
  • MCP integration to connect external tools via SSE, HTTP, or WebSockets.
  • Smart UI features including a side panel, floating selection toolbar, and image/screenshot input.
  • Context management through summary compression and message re-editing for API providers.