L

Lightpanda Browser

SHA-256
36.9k stars·AI Productivity·SHA-256 checksum verified

A from-scratch headless browser in Zig that uses about 16x less memory and runs roughly 9x faster than headless Chrome, designed for AI agents and web automation.

Smart Download

Download Download Version

v1.0.1 · 69.8 MB

An ultra-lightweight headless browser purpose-built for AI agents and automation.

Core Features

  • Ground-up browser engine written in Zig, not a Chromium fork
  • Roughly 16x lower memory usage and 9x faster execution than headless Chrome
  • Speaks CDP and WebDriver BiDi, working as a drop-in backend for Puppeteer and Playwright
  • Built-in agent mode for driving the browser in plain English with major LLM providers or local models
  • Native MCP server giving AI clients like Claude direct web browsing tools

What It Can't Do

  • •There is no native Windows build, so Windows users must run it inside WSL2. Linux binaries are linked against glibc and will not run on musl-based distributions like Alpine without building from source. As a brand-new engine, compatibility with heavily JavaScript-dependent pages may not match real Chrome, so test your target sites before production use. The AGPL-3.0 license has copyleft requirements that may affect closed-source commercial integrations.

Use Cases

  • Large-scale web browsing, clicking, and data extraction for AI agents
  • Low-cost replacement for headless Chrome in scraping and automation pipelines
  • Giving LLM applications web access through the MCP protocol
  • Web automation on memory-constrained servers and containers

Detailed Introduction

Lightpanda is an open-source headless browser written from scratch in Zig, purpose-built for AI agents and web automation. Unlike Chromium-based solutions such as headless Chrome driven by Puppeteer or Playwright, it is not a fork or patch of an existing engine but a clean-room implementation. Benchmarks over 933 real web pages show roughly 16x lower peak memory (123MB vs 2GB) and about 9x faster execution than headless Chrome, which makes large-scale scraping and AI browsing workloads dramatically cheaper to run. Despite being a brand-new engine, it speaks the Chrome DevTools Protocol and WebDriver BiDi, so existing Puppeteer and Playwright scripts work with minimal changes. A built-in agent mode lets you drive the browser in plain English using Anthropic, OpenAI, Gemini, OpenRouter, or local models via Ollama, and export deterministic, token-free replayable scripts. A native MCP server exposes browsing tools to AI clients like Claude. Licensed under AGPL-3.0.

Troubleshooting & FAQ (1)

Troubleshooting
Why does fetch() reject with 'TypeError: fetch error' on HTTP 401 in Lightpanda when request interception is enabled?

This is a Lightpanda bug (present in 1.0.1 and main @3476970a). With Fetch.enable({handleAuthRequests: true}) — which is what Puppeteer's page.setRequestInterception(true) sends — Lightpanda treats every 401/407 response as an HTTP auth challenge even when the response has no WWW-Authenticate/Proxy-Authenticate header. It emits Fetch.authRequired with an empty challenge (source: "", scheme: ""), and when the client replies with Default or CancelAuth (Puppeteer's default), Lightpanda aborts the request with AbortAuthChallenge instead of delivering the 401 to the page. SPAs using cookie/session auth that return bare 401s therefore see fetch() reject. Workarounds: (1) Enable interception via raw CDP with handleAuthRequests: false — the http_request_auth_required listener is then never registered, the transfer is not parked, and the 401 reaches the page. (2) Scope Fetch.enable with URL patterns that exclude your API, e.g. Fetch.enable({patterns: [{urlPattern: '*/assets/*'}], handleAuthRequests: true}) — requests not matching the pattern bypass interception and receive the real 401. Chrome differs in two ways: it only fires Fetch.authRequired when a real challenge header exists, and for Default/CancelAuth answers it cancels the auth flow but still delivers the original 401 response. Track issue #4018 for a fix aligning Lightpanda with Chrome.

GitHub Issue #4018

Tags

headless-browserai-agentweb-automationweb-scrapingzigmcppuppeteer-alternative

Getting Started

1

Download installer

Click the button above to download the installer for your system

2

Install the software

Install the appropriate package for your distro (dpkg / rpm / AppImage)

3

Install via Homebrew with brew install lightpanda-io/browser/lightpanda, or download the binary for your platform from the GitHub nightly releases

4

Run lightpanda version in your terminal to verify the installation

5

Fetch a page instantly with lightpanda fetch --dump html URL

Install Guide
  1. Install via Homebrew with brew install lightpanda-io/browser/lightpanda, or download the binary for your platform from the GitHub nightly releases
  2. Run lightpanda version in your terminal to verify the installation
  3. Fetch a page instantly with lightpanda fetch --dump html URL
File Integrity

SHA-256 checksum verified

Checksum extracted from GitHub official Release page

SHA256 Checksum

1fcca0ae7684f470780de611a56e9a8ea612d5ff21e37dbb48905ebac5b918a7

This checksum is extracted from the GitHub Release page. Verify file integrity after download.

All SHA-256 checksums on this platform are extracted from the project's official GitHub Release page, without any modification. You can independently verify them on the GitHub Releases page.

Open Source Transparency

View GitHub Source
Environment Guide

Uninstall Info

If installed via Homebrew, run brew uninstall lightpanda. If you downloaded the binary manually, simply delete the file. For Docker, remove the container and image with docker rm lightpanda followed by docker rmi lightpanda/browser. No other files are left behind on the system.

No Extra Dependencies

Ready to use after download. No additional runtime required.

Project Info
LicenseAGPL-3.0
Last Updated2026-10-23 10:17:42
GitHub RepositoryOfficial Website

Having issues? Check the FAQ below

1 FAQ

Similar Projects