yeahhe365/Gemini-Nexus

What it solves

Gemini Nexus is a Chrome extension that integrates multiple AI providers into a single browser interface. It eliminates the need to switch between different AI web tabs or manage multiple API keys across different platforms by providing a unified side panel and floating toolbar for AI-assisted browsing.

How it works

The extension acts as a native AI layer for the browser, supporting various connection methods:

• Provider Integration: It connects to Gemini Web (via reverse-engineered RPC endpoints), the official Gemini API, OpenAI-compatible APIs (including DeepSeek, OpenRouter, and Qwen), Anthropic, and Zhipu.
• Browser Control: Using the Chrome DevTools Protocol, it allows the AI to perform agentic tasks like clicking, typing, navigating pages, and taking snapshots of the page content.
• External Tooling: It supports the Model Context Protocol (MCP) via SSE, WebSocket, or HTTP, allowing the AI to access tools from external MCP servers through a local proxy.
• UI Integration: It injects a floating toolbar for text selection actions (translate, summarize) and a side panel for full conversations.

Who it’s for

• Users who want a single interface to access multiple LLMs while browsing.
• Developers or power users who want to automate browser tasks using AI agents.
• People looking to integrate external MCP tools directly into their browser workflow.

Highlights

• Multi-Provider Support: Seamlessly switch between Gemini Web, official APIs, and various third-party providers.
• Agentic Browser Control: Ability to perform real browser actions like filling forms and managing tabs.
• MCP Integration: Connects to external MCP servers to extend the AI's capabilities.
• Advanced Input: Supports OCR, screenshot translation, and screen/window capture for image-based prompts.
• Context Management: Includes summary compression and recent-turn trimming to handle long conversations.