A guide to building long-term compounding knowledge infrastructure. See details on GitHub .

Nanobrowser

Open-source in-browser AI web automation tool with multi-agent support and flexible model options, focused on privacy and local execution.

Introduction

Nanobrowser is an open-source Chrome extension for AI-powered web automation that runs in the browser. It enables multi-agent workflows and lets users connect their own LLM API keys for flexible, privacy-preserving automation, positioning itself as a free alternative to centralized operator products.

Key features

  • Multi-agent system: Specialized agents collaborate to perform complex web workflows.
  • Multiple LLM support: Works with OpenAI, Anthropic, Gemini, Ollama and other providers.
  • Local execution & privacy: Core automation runs in the browser; users keep control of API keys and data flow.
  • Interactive side panel: Conversational UI with real-time status and task control for easy debugging.

Use cases

  • News summarization and monitoring: scrape and summarize headlines from target sites.
  • GitHub and research automation: bulk search and extract repository insights.
  • Repetitive form filling and data entry: automate tedious web tasks.
  • End-to-end sampling and testing: run workflows in a real browser environment for reliable sampling.

Technical highlights

  • Built as a Chrome extension with full support for Chrome and Edge feature sets.
  • Developed in TypeScript inside a monorepo layout, separating frontend and agent subsystems.
  • Optional local model support (e.g., Ollama) enables offline inference and lower operational cost.

Comments

Nanobrowser
Resource Info
Author Nanobrowser
Added Date 2025-09-18
Tags
OSS AI Agent Dev Tools