Nanobrowser

Nanobrowser is a Chrome extension that automates web tasks with a multi-agent system running inside your browser. You describe a task in a side panel chat, and specialized agents (a Planner and a Navigator) work through it, with the Planner adjusting instructions when the Navigator meets obstacles. The extension uses your own LLM API keys, so credentials stay local and you pay providers directly.

Supported providers include OpenAI, Anthropic, Gemini, Ollama, Groq, Cerebras, Llama, and custom OpenAI-compatible endpoints, and you can assign a different model to each agent, for example a larger model for planning and a smaller one for navigation. Local models through Ollama or another OpenAI-compatible server are supported, with community-tested models such as Qwen3-30B-A3B-Instruct-2507, Falcon3 10B, Qwen 2.5 Coder 14B, and Mistral Small 24B.

Chrome and Edge are officially supported; Firefox, Safari, and other Chromium variants are not. Example tasks include extracting headlines from a news site, finding trending repositories on GitHub, and product searches with constraints.

Features

• Multi-agent system: Planner and Navigator agents collaborate on multi-step web workflows
• Side panel chat: task input with real-time status updates and follow-up questions about completed tasks
• Per-agent model choice: different providers or models for the Planner and the Navigator
• Provider support: OpenAI, Anthropic, Gemini, Ollama, Groq, Cerebras, Llama, and OpenAI-compatible endpoints
• Local models: run through Ollama or a custom endpoint with no data leaving the machine
• Conversation history: stored interaction history you can revisit and manage
• Browser support: Chrome and Edge, installed from the Chrome Web Store or as an unpacked build