#163163rd of 389 in Skills, plugins and prompts
lcu
Exposes the ChatGPT app's computer-use runtime to Claude Code, Codex CLI and Pi
- Stars
- 2.4k
- License
- MIT
- Last commit
- Nov 2026
- Last release
- Nov 2026
- Language
- JavaScript
1 on 22 Sept 20262,412 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
LCU lets coding-agent harnesses drive desktop apps (read windows, click, type, screenshot) and Chrome tabs by reusing the computer-use runtime from the locally installed ChatGPT desktop app, without Codex authentication. It ships adapters for Pi, Codex CLI and Claude Code, plus experimental ones for Oh My Pi, Hermes, DeepSeek Harness and Junie. Per-app approvals are handled by LCU, including a native approval pane in Claude Code and Touch ID-gated `lcu apps` commands on macOS.
Who it is for: Developers adding desktop and Chrome control to Claude Code, Codex CLI or Pi
Strengths
- Adapters for Claude Code, Codex CLI and Pi, with experimental ones for four more harnesses
- Per-app approval stays enforced in every permission mode, with Touch ID needed to allow or revoke on macOS
- Needs no Python; runs on the Node bundled in the ChatGPT app
- Release archives ship with SHA-256 checksums for macOS arm64, Linux arm64 and Linux x64
Weaknesses
- Requires the official ChatGPT desktop app installed locally; it supplies the runtime
- Intel Macs, native Wayland and musl Linux are unsupported; Windows 11 is only a candidate
- Claude Code Chrome support is experimental; some interruptions skip tab cleanup
- Junie approvals are permanent until revoked because Junie cannot show the per-app prompt
What it needs
- Needs ChatGPT desktop app, Claude Code, Codex CLI, Pi or another supported harness, X11 and D-Bus (Linux), Chrome with the ChatGPT extension (optional)
Quick answers
Is lcu open source?
Yes. Its code is on GitHub under the MIT license.
Also in Skills, plugins and prompts
See all 389| Rank | Project | Score |
|---|---|---|
| 1 | claude-memPersistent cross-session memory plugin for Claude Code and other coding agents | 97 out of 100 |
| 2 | impeccableDesign skill with 24 slash commands for AI coding agents' frontend output | 97 out of 100 |
| 3 | gstackMarkdown slash-command skills that give Claude Code planning, review, QA and ship roles | 96 out of 100 |
| #4 | rufloAgent harness adding swarms, memory and hooks to Claude Code and Codex | 96 out of 100 |
| #5 | cavemanShortens coding-agent replies and compresses tool output before the model reads it | 95 out of 100 |