What can I use browser for?

Pass a news article URL to the dump command and get its readable text back as a markdown file; Point an existing Puppeteer scraping script at Lightpanda's CDP server and collect the same page data with less memory; Tell agent mode to fill in a web form, then save the session as a script you can rerun without an AI model

Can I run browser on my computer?

It runs on your own computer or server after a local setup. No native Windows or Android build, and the Linux binary fails on musl systems such as Alpine.

What do I need to use browser?

Linux (glibc-based) or macOS, on x86_64 or aarch64, or Docker. Windows works only inside WSL. Agent mode needs a key for an AI provider or a local model.

browser: Load and control web pages for AI agents with less memory

What it does

Lightpanda is a headless browser (one that loads and runs web pages without showing a window), written from scratch in the Zig language rather than built on Chrome. It parses HTML, runs JavaScript and handles clicks, forms and cookies. In the creators' benchmark of 933 real pages on one cloud server, peak memory for 100 pages was 123MB against 2GB for headless Chrome, and the run took 5s against 46s.

How to use it

Input → action → output: give it a URL with the dump command and get the page back as markdown, structured JSON metadata, or a file. Developers can instead start a CDP server (the protocol Puppeteer and Playwright use to control Chrome) on port 9222 and point existing scripts at it. In agent mode you describe a task in plain English, and it navigates, fills forms and extracts data, then saves the steps as a replayable JavaScript file.

Common use cases

Pass a news article URL to the dump command and get its readable text back as a markdown file.
Point an existing Puppeteer scraping script at Lightpanda's CDP server and collect the same page data with less memory.
Tell agent mode to fill in a web form, then save the session as a script you can rerun without an AI model.

Can I use this?

Runs in Docker on your computer · needs an AI model key

Requirements Install the nightly build with Homebrew, the Arch User Repository, a downloaded binary, or the official Docker image, which exposes the CDP server on port 9222. It also runs as an MCP server (a way to plug tools into an AI assistant). AGPL-3.0 licensed, so it is open source.

Caveat Benchmark numbers come from the creators' own crawler test; sites that rely on browser features Lightpanda has not implemented may behave differently.

Worth knowing No native Windows or Android build, and the Linux binary fails on musl systems such as Alpine. Usage telemetry is on by default; an environment variable turns it off.

Why it's here

2,635 people starred it on GitHub this week — 41,593 in total.

Numbers as of 8 November 2026, from BuildTube’s latest daily refresh.

Behind it

View setup guide

BuildTube has not run or verified this project. Everything above is written from what the creator published.

Made this?

Quick questions

What is browser?: Lightpanda is a headless browser (one that loads and runs web pages without showing a window), written from scratch in the Zig language rather than built on Chrome.
What can I use browser for?: Pass a news article URL to the dump command and get its readable text back as a markdown file; Point an existing Puppeteer scraping script at Lightpanda's CDP server and collect the same page data with less memory; Tell agent mode to fill in a web form, then save the session as a script you can rerun without an AI model
Can I run browser on my computer?: It runs on your own computer or server after a local setup. No native Windows or Android build, and the Linux binary fails on musl systems such as Alpine.
What do I need to use browser?: Linux (glibc-based) or macOS, on x86_64 or aarch64, or Docker. Windows works only inside WSL. Agent mode needs a key for an AI provider or a local model.
What license does browser use?: AGPL-3.0 licensed, so it is open source.

More from this week

OpenBot Run AI coworkers with their own browser and files
Give an AI coworker a job, connect the tools it needs and watch it work.
+1,312 stars this weekDesktop appAdded 7 Nov
agent-desktop Let AI agents read and click desktop apps
For developers building computer-use agents: agent-desktop lets an AI agent read and operate desktop apps on Mac, Windows or Linux from a command line.
+13,858 stars this weekAdded 3 Nov
Auto-Ontology Give AI agents checked business meanings for your database
Help AI agents answer questions about company databases using reviewed definitions and SQL, with each answer showing its steps.
+9,044 stars this weekNeeds a local setupAdded 4 Nov

Related projects

Moli Extract web pages for AI agents with Moli
Same topic: browser
Added 12 Oct
ego-lite Let AI agents browse in their own tabs while you keep working
Same topic: browser
Added 21 Oct
Scrapling Collect data from websites with Python
Same topic: playwright
+8,712 stars this weekNeeds a local setupAdded 1 Nov
OpenBot Run AI coworkers with their own browser and files
Same topic: browser-automation
+1,312 stars this weekDesktop appAdded 7 Nov
Playwright Let tests, scripts or AI agents click through websites for you
Same topic: playwright
+627 stars this weekAdded 21 Oct
ai-goofish-monitor Watch Xianyu listings and get alerts on good finds
Same topic: playwright
Needs a local setupAdded 4 Nov

Explore by topic

Post this

For makers

Made this? Show it’s on BuildTube — add the badge to your README or embed the card on your site.

README badge (Markdown)

[![Featured on BuildTube](https://buildtube.app/badge/gh-lightpanda-io-browser.svg)](https://buildtube.app/p/gh-lightpanda-io-browser/)

Embed card (HTML)

<iframe src="https://buildtube.app/embed/gh-lightpanda-io-browser/" width="560" height="240" style="border:0;max-width:100%" loading="lazy" title="browser on BuildTube"></iframe>