Nanobrowser

Use it when

• You want to automate repetitive multi-step web tasks by typing a natural-language instruction into your own browser instead of writing a script
• You want an OpenAI Operator style browser agent that is free, runs locally in Chrome or Edge, and uses your own LLM API keys or local Ollama models

What it solves

Not the fit when

• Chrome and Edge are the only officially supported browsers; Firefox, Safari, Opera, Arc, and other Chromium variants are not supported
• Remote-model configurations require your own provider API keys and incur that provider's usage charges
• The extension runs in your local browser, but tasks sent to OpenAI, Anthropic, Gemini, Groq, Cerebras, or another remote provider cross that provider's data boundary; use local Ollama when data must not leave the machine
• Local models need more explicit step-by-step prompts and may be less stable on complex tasks
• Not a headless server-side scraper, CI browser API, or scripted Playwright/Puppeteer replacement
• headless server-side scraping
• programmatic browser automation for CI pipelines
• Firefox or Safari automation
• scriptable API-driven page control

Install

Add to Chrome from the Nanobrowser Chrome Web Store page (or download nanobrowser.zip from GitHub releases and Load unpacked at chrome://extensions with Developer mode on)

Invoke

Click the Nanobrowser toolbar icon to open the side panel, add LLM API keys in Settings and assign models to the Planner and Navigator agents, then type a task like 'Go to TechCrunch and extract the top 10 headlines from the last 24 hours'

Alternatives

No reviewed alternatives recorded yet.