ArchestackBest of Open-Source AI: List on GitHub ↗3 stars (opens in a new tab)

#2929th of 48 in Agent frameworks and SDKs

nanobrowser

Browser agent in a Chrome/Edge side panel that clicks and types for you

Website ↗ (opens in a new tab)Repository on GitHub ↗ (opens in a new tab)
Stars
14k
License
Apache-2.0
Last commit
Oct 2026
Last release
Oct 2026
Language
TypeScript

Missed the latest GitHub check; these facts may be older.

Overview

Nanobrowser is a Chrome and Edge extension. You give a task in a side panel, and an LLM agent reads pages, clicks, types and switches tabs. It uses your own API key or a local model (Ollama, LM Studio, OpenAI-compatible servers). In the default Ask mode it asks before submit, send, delete or upload, and per-site Allow, Ask and Block rules are set in Settings.

Who it is for: Users who want a browser-driving agent with their own API keys or local models

Strengths

  • Per-site Allow/Ask/Block rules; Ask mode confirms submit, send, delete and upload
  • Supports Anthropic, OpenAI, Gemini, OpenRouter, xAI, DeepSeek, Mistral, Ollama and LM Studio
  • Several tasks run in parallel, each in its own tab group
  • Deterministic control layer on chrome.debugger with element IDs tied to exact elements

Weaknesses

  • Only Chrome and Edge are supported; other browsers are untested
  • Models must support tool calling; local models suit only simple tasks
  • Local models need 16k+ context plus Ollama origin or LM Studio CORS setup
  • README says prompt-injection protection is imperfect and the agent can miss steps needing approval

What it needs

  • GPU optional
  • Needs Chrome or Edge, LLM provider API key or local model server, Node.js 24 and pnpm 10 (build from source)
  • Models: Anthropic, OpenAI, Google Gemini, OpenRouter, xAI Grok

Also in Agent frameworks and SDKs

See all 48
Also in Agent frameworks and SDKs
RankProjectScore
1ppt-masterAgent skill that turns documents into natively editable PowerPoint decks Live demo ↗ (opens in a new tab)62.1k starsMIT93 out of 100
2browser-usePython library that lets an LLM agent control a web browser Live demo ↗ (opens in a new tab)118k starsMIT89 out of 100
3langchainPython framework for building LLM agents and apps from interoperable components147.8k starsMIT88 out of 100
#4tradingagentsLangGraph multi-agent framework simulating a trading firm to analyze tickers Live demo ↗ (opens in a new tab)111.2k starsApache-2.088 out of 100
#5copilotkitSDK for adding agent chat, generative UI and shared state to apps38k starsMIT87 out of 100