#2626th of 118 in Agent frameworks and SDKs
ego-lite
Desktop browser that AI agents drive through isolated per-agent Spaces
- Stars
- 22.5k
- License
- MIT
- Last commit
- Nov 2026
- Last release
- Nov 2026
- Language
- JavaScript
74 on 21 Apr 202622,532 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
ego lite is a Chromium-based desktop browser for macOS and Windows that gives each AI agent its own isolated Space, so agents work in the background while you keep browsing. Agents such as Claude Code, Codex or Cursor connect through the `ego-browser` skill, which exposes snapshot, fill, click, wait, navigate and capture as in-page JavaScript functions. The repository holds the MIT-licensed skill and docs; the browser itself is a separate free download.
Who it is for: Developers running Claude Code or Codex agents on web automation tasks
Strengths
- Per-agent Spaces let several agents run in parallel without taking over your tabs
- Agent writes one JavaScript snippet per multi-step task, not a command-by-command loop
- Optional Chrome data migration gives agents your existing logins, cookies and extensions
- Works with any agent CLI that supports skills, installable via npx skills add
Weaknesses
- Runs on macOS and Windows only; Linux is just on the roadmap
- Browser binary is a separate download, not built from this MIT repository
- Benchmark of 31 tasks is self-run by the authors against four other tools
- Experience-accumulation reusable skills feature is marked coming soon
What it needs
- Needs Agent CLI (Claude Code, Codex, Cursor or custom), ego lite desktop browser app
- Models: Model-agnostic; depends on the agent CLI
Quick answers
Is ego-lite open source?
Yes. Its code is on GitHub under the MIT license.
Also in Agent frameworks and SDKs
See all 118| Rank | Project | Score |
|---|---|---|
| 1 | ppt-masterAgent skill that turns documents into natively editable PowerPoint decks | 95 out of 100 |
| 2 | openai-agents-pythonPython SDK for multi-agent workflows with handoffs, guardrails and tracing | 91 out of 100 |
| 3 | mastraTypeScript framework for building AI agents, workflows and MCP servers | 91 out of 100 |
| #4 | browser-usePython library that lets an LLM agent control a web browser | 90 out of 100 |
| #5 | copilotkitSDK for adding agent chat, generative UI and shared state to apps | 90 out of 100 |