Free Intensive Nov 21–22Free 2-day hands-on Vibecoding Intensive on Nov 21–22 at 16:00 MSK (Sat–Sun) · Live Cohort 3 starts Nov 23Join Free →Cohort 3 (Nov 23)

Ego lite Gives Each AI Agent Its Own Workspace Space in the Browser

Usually, an AI agent takes over your Chrome, and while it clicks through pages, you sit and wait. Ego lite is a free browser for Mac and Windows that gives each agent its own Space: a separate workspace inside the same window. Claude Code, Codex, or Cursor run there in the background under your logins, while you keep working in your own tabs. For a setup of "ten leads, ten Spaces," this is more convenient than handing the agent a single browser.

AI-processed from TG: @TochkiNadAI (Точки над ИИ); edited by Hamidun News
Ego lite Gives Each AI Agent Its Own Workspace Space in the Browser
Source: TG: @TochkiNadAI (Точки над ИИ). Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceTG: @TochkiNadAI (Точки над ИИ) ↗
Editor-in-Chief
Key metric #1Free, for macOS and Windows, with data staying on the device
Key metric #2On first launch, it offers to import cookies and extensions from Chrome
Key metric #3You can peek at any agent or take over its control

Imagine you asked an AI agent to gather data on ten companies. It takes over your Chrome and starts clicking through pages, while you sit and wait, because you and the agent share one browser. Ego lite is a free browser for Mac and Windows that tries to eliminate exactly this problem. The core idea: each agent (Claude Code, Codex, or Cursor) gets its own Space, a separate workspace inside the same browser. The agent works there in the background under your logins, while you keep using your regular tabs.

How the browser is built for humans and agents

AI browsers are usually built in one of two ways. The first embeds the agent directly into the browser, as Comet does. The second uses an extension through which Claude Code or GPT controls your Chrome. In both cases, while the agent works, you get in its way, and it gets in yours.

Ego lite was designed from the start to be shared by you and your agents. A Space is like a separate desk in a shared office. The agent sits at its desk, opens pages, and does its work, while you sit at yours. You can see what it is doing, but you don't have to wait for it to finish.

The agent works under your logins. That is why, on first launch, ego lite offers to import cookies and extensions from Chrome. This removes the most tedious step: you don't need to log into your CRM, email, and social networks again so the agent can access them.

Ten agents, ten tasks

The main scenario in the description is parallel work. You can launch a batch of agents at once, for example ten Spaces for ten leads. Each agent handles its own lead: it opens the company's website, looks for contacts, and fills in the record. At any moment you can switch to a particular Space, see what the agent is doing, or take over and finish a step by hand.

That last point matters in practice. Agents in a browser often run into a CAPTCHA, two-factor authentication, or an unusual form. If there is only one browser for everyone, a single snag halts everything. If each agent has its own Space, a stuck agent doesn't block the other nine, and you step in only where one has stalled.

The technical details are few. Ego lite runs on macOS and Windows, is free, and keeps data on the device. The code and description are on GitHub in the citrolabs/ego-lite repository.

What remains open

The author of the original post honestly admits he hasn't yet figured out why he needs it. The objections are clear. There are already plenty of agents in browsers. IDEs have had built-in browsers for a long time. There is browser use and dozens of skills that teach an agent to navigate websites. Ego lite will have to prove that a separate application is better than this existing set of tools.

There is also a security question that the description does not address. You hand agents your cookies and extensions, which means access to your work accounts. If an agent makes a mistake or lands on a malicious page with a prompt injection, it will act on your behalf. Before importing everything from Chrome, it is better to set up a separate profile with restricted permissions or separate accounts for the agents. The post doesn't explain how ego lite isolates Spaces from one another, and that should be verified before any serious use.

Who should try it

The project is interesting to people who already run several agents in parallel and are tired of sharing one Chrome among them. Typical tasks include gathering data from a list of companies, checking dozens of pages, and performing routine actions in web interfaces that have no API. If you use agents a couple of times a week, the gain will be small.

The takeaway for developers and product managers is this. The next step for agentic browsers is not just a smarter model, but a more convenient division of workspace between a human and several agents. A separate space for each agent, and the ability to peek in and take over control, are features that will likely show up in other products too. Ego lite is free, so testing it on a couple of real tasks costs almost nothing. But the logins you hand over to it are expensive.

Story

Key Questions Answered

How the browser is built for humans and agents?

AI browsers are usually built in one of two ways. The first embeds the agent directly into the browser, as Comet does. The second uses an extension through which Claude Code or GPT controls your Chrome. In both cases, while the agent works, you get in its way, and it gets in yours. Ego lite was designed from the start to be shared by you and your agents. A Space is like a separate desk in a shared office. The agent sits at its desk, opens pages, and does its work, while you sit at yours.

What are the key details on Ten agents, ten tasks?

The main scenario in the description is parallel work. You can launch a batch of agents at once, for example ten Spaces for ten leads. Each agent handles its own lead: it opens the company's website, looks for contacts, and fills in the record. At any moment you can switch to a particular Space, see what the agent is doing, or take over and finish a step by hand. That last point matters in practice. Agents in a browser often run into a CAPTCHA, two-factor authentication, or an unusual form.

What remains open?

The author of the original post honestly admits he hasn't yet figured out why he needs it. The objections are clear. There are already plenty of agents in browsers. IDEs have had built-in browsers for a long time. There is browser use and dozens of skills that teach an agent to navigate websites. Ego lite will have to prove that a separate application is better than this existing set of tools. There is also a security question that the description does not address.

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Hands-On AI Engineering · Hamidun Academy

Build & Ship Production AI Agents Without Boilerplate Code

Start your free 72-hour trial at Hamidun Academy (academy.hamidun.com) — interactive lessons on Claude Code & Cursor, live Tech Radar, and production agent templates.

Start 72h Free Trial →Vibe-Coding Intensive
What do you think?
Loading comments…