libi

libi / ko-browser

Public

A simple, fast, token-efficient browser for AI agents

agent agent-browser ai-browser-automation ai-browser-control browser
12
0
100% credibility
Found Apr 03, 2026 at 12 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Go
AI Summary

ko-browser is a lightweight browser control tool for AI agents that delivers compact, token-saving page snapshots via simple command-line instructions or embeddable code.

How It Works

1
💡 Discover ko-browser

You hear about a simple tool that lets AI helpers browse websites like a human, using easy numbered lists instead of confusing pictures.

2
📥 Get it set up

Download and install with one quick command, and it grabs a web browser if needed so everything is ready to go.

3
🌐 Open a website

Tell it to visit any page, like Google or your favorite site, and it loads right up.

4
👁️ See the page clearly

Snap a quick view of the page as a neat numbered list of buttons, links, and boxes – super simple for your AI to understand without wasting space.

5
🖱️ Click and type easily

Use numbers like 'click 3' or 'type hello in 5' to interact, just like pointing at things on screen.

6
⏳ Wait and repeat

Let it pause for pages to load, snap new views, or grab pictures – your AI keeps exploring smoothly.

✅ AI browses perfectly

Your AI agent now navigates any website fast and smart, understanding everything with tiny, efficient descriptions that save time and effort.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 12 to 12 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is ko-browser?

ko-browser is a Go-based browser automation tool for AI agents, delivering a CLI with 86 commands and an embeddable library to control Chrome via its DevTools Protocol. It generates compact accessibility tree snapshots that cut token usage by 46% compared to verbose alternatives, making it ideal for agent loops like open, snapshot, click, type, and wait. Users get a single binary—no Node.js or Playwright runtime—for tasks like form filling, screenshots, and OCR on image-heavy pages.

Why is it gaining traction?

It stands out with ~50ms startup versus 500ms for Node tools, numeric element refs like "click 5" post-snapshot, and dual CLI/library use for shell agents or Go apps. Optional Tesseract OCR handles non-DOM text, while features like network blocking, state export, and annotated screenshots solve real agent pains without bloat. Devs dig the simple GitHub action workflow examples for CI automation.

Who should use this?

AI agent builders scripting browser flows in shell or embedding in Go runtimes. Automation scripters needing chrome browser ko kaise kholen, browser ko kaise hataye, or browser ko samjhaie without heavy deps. Go backend teams handling simple fasting app reviews or simple fast loans scraping.

Verdict

Promising for niche AI browser needs, but 12 stars and 1.0% credibility signal early maturity—docs are solid, but expect rough edges in edge cases. Try for token-thrifty prototypes; skip for production without more testing.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.