The best browser for both you and your AI agents
An agent browser built for AI agents to run many browser automation tasks in parallel, much faster and cheaper every run.
Parallel multitasking,
execute faster.
ego (lite) runs multiple in-page actions in a few lines of JavaScript, instead of one tool call at a time. Complex browser tasks finish up to 245% faster, on fewer tokens.
AI agents can also smoothly multitask many browser automation tasks at once, each in its own Space, without colliding.
Same automation, fewer tokens
ego (lite) 's semantic snapshot is built inside our custom Chromium engine, not a JS shim on stock Chrome. Snapshots come back faster, ref actions run more reliably, and reach into cross-origin iframes, shadow DOM, and third-party SDK widgets like Stripe, Salesforce, Intercom, and React portals.
Web fetch and JS shims usually give up on those. Your AI agent ends up retrying against content that was never really there, burning tokens for nothing.
Space for agents, Space for you
Spaces are workspaces both you and your agent can use. Group your tabs by project in yours; Your agent runs tasks on its own. Giving you and your agent separate workspaces in the same browser.
Same Chrome, agent-native
Built on Chromium, ego (lite) imports your entire Chrome setup in one click. Use it as your default browser, and your AI agents will never get stuck on captchas, 2FA, and SSO redirects.
Finishing more browser automation tasks, faster, and at lower tokens cost than any blank-Chromium agent.
Any agent, connect via
ego-browser
Drop-in ego-browser ships with built-in skills for ego, Claude Code
, Codex
, Kiro
, Hermes agent
, OpenClaw
, Cursor
and any other AI agent that writes codes.
You can even let many different AI agents drive ego (lite) at the same time, each working in its own Space, without ever interrupting yours.
Stop bouncing
between browsers
ego (lite) replaces your daily Chrome, your AI agent's Chromium runtime, and the SDK glue them together.
Pick the right one
| Feature | ego (lite) | Browser Use | Agent Browser | ChatGPT Atlas | Perplexity Comet |
|---|---|---|---|---|---|
| Multitask | ✓ | — | — | — | — |
| Reusable skills (coming soon) | ✓ | — | — | — | — |
| Inherits Chrome's data | ✓ | — | — | ✓ | ✓ |
| Same browser, separate workspace | ✓ | — | — | — | — |
| Compressed semantic input | ✓ | — | ✓ | — | — |
| Controllable by external agents | ✓ | ✓ | ✓ | — | — |
| Data stored locally | ✓ | ✓ | ✓ | — | — |
| No login friction | ✓ | — | — | ✓ | ✓ |
| Daily-use browser | ✓ | — | — | ✓ | ✓ |
| Free, no subscription required | ✓ | ✓ | ✓ | — | — |
FAQ
ego lite is a Chromium browser built for you and your AI agent to work together, making browser automation faster and cheaper, every run.
Existing tools like browser-use and agent-browser are browser automation frameworks. They need a separate browser to drive, logins never carry cleanly, and you and the agent end up fighting for the same tabs.
ego lite is one browser designed from the start for the two of you to share — no extra setup, and the agent can always reach your real logins and tabs through ego-browser.
Playwright and Puppeteer are browser automation libraries you embed in code. They drive pages through CSS selectors, which break whenever a site reshuffles its DOM.
ego lite is a real browser plus ego-browser. ego-browser is the connection layer between any agent CLI (Claude Code, Codex, Cursor, or your own) and ego lite — it exposes the browser as a set of in-page JavaScript tools: snapshot, fill, click, wait, navigate, capture. The agent writes a JavaScript snippet that composes several steps using those tools, and ego-browser runs it on the page in one pass, instead of the one-tool-at-a-time round trips Playwright and Puppeteer require. The snapshots cost fewer tokens than HTML, and the @N refs stay stable when class names rotate. There's no SDK to integrate, and nothing to install from the Playwright or Puppeteer ecosystem.
ego lite is a real browser plus ego-browser. Whichever agent you use, whether that's ego, Claude Code, Codex, Cursor, or something else, drives the same browser you use daily. Space keeps the agent's work isolated from yours.
Browser-Use is an autonomous agent framework. It bundles its own LLM loop and drives a separate browser instance. You can point it at a real Chrome, but only after you fully close it. Extensions don't load by default. And it has its own pricing tier on top of whatever you already pay for.
Vercel's agent-browser is a CLI that lets your agent control a browser through shell commands, but it ships no browser of its own. The agent doesn't share your daily browser session, so your real logins aren't available to the running task.
Yes. Any agent that can run a shell command can drive ego lite through ego-browser:
- ego, Claude Code, OpenAI Codex, Cursor, Continue, Gemini CLI
- Hermes Agent, OpenClaw, Opencode
- Any custom in-house agent you've built
ego-browser installs as a skill into your agent's skill directory during onboarding. The agent then calls browser automation commands the same way it runs any other shell tool. There's no SDK to integrate, and you're not locked into a specific model.
Yes. ego lite is built on Chromium, so the address bar, tab groups, bookmarks, profiles, extensions, downloads, and incognito mode behave the way they do in Chrome. The difference is Space: an isolated workspace where your AI agent runs web automation alongside you, and doesn't touch your tabs unless you let it.
Yes. ego lite's onboarding migrates the following from Chrome in one click:
- Tabs and tab groups you're using
- Bookmarks and saved passwords
- Browser extensions
- Cookies and login sessions
- Browser profiles
Every site you've already signed into stays signed in, and the agent inherits the same logged-in state. It doesn't need to re-authenticate to Gmail, Notion, LinkedIn, or your internal tools every time it runs a task.
Yes. The agent runs inside Space, an isolated workspace in ego lite. It works in its own area, so your tabs stay untouched unless you explicitly hand one over. It doesn't take over your mouse or steal focus from whatever you're doing. Switch into Space anytime to watch the agent work, or ignore it and keep working in your own window.
What gets cheaper is the whole task, not any single step. The savings come from several features stacking together.
The agent can run several actions in the page with a few lines of JavaScript, instead of calling one tool at a time and waiting for each result before deciding the next, the way Playwright, Puppeteer, and most browser automation frameworks work.
Combined, complex browser automation tasks complete faster and on fewer tokens overall.
Yes. ego lite doesn't upload your browsing data. Your history, cookies, login sessions, and anything the agent reads stay on your own computer. The one thing ego lite records during setup is whether you opted into Chrome data migration.
ego (lite) doesn't collect any data
Bookmarks, browsing history, cookies, passwords, every page you load — all stored on your computer, we can't access them.