About LoopX
LoopX is a Mac app that runs AI agents on a schedule, so routine triage, status checks and first drafts stop costing your best hours.
Last updated:
01
Why it exists
Engineering days get eaten by small, repetitive work: issues that sit untouched, an inbox that buries the urgent mail, pipelines that fail while you are in a meeting. None of it needs your best thinking, yet all of it needs doing.
LoopX takes that work on end to end and hands the decision back to you at the point where judgment is needed. The mission is simple: give engineers their attention back.
02
What it does
LoopX organizes work into loops. Each loop is an independent routine with its own schedule, instructions and run history, for example fixing GitLab issues, triaging your inbox, preparing for meetings or writing release notes.
A local dashboard shows every run, what it did and why, and an Office view lets you talk to your agents in a 3D office.
03
Principles we keep
These are design rules, not slogans. Each one is visible in how the app behaves.
- Local first. LoopX runs on your Mac with your own tools and credentials. Nothing is hosted for you.
- Human in the loop. It stops at an open merge request or a draft. It never merges and never sends email.
- Bring your own agent. LoopX ships no model and sells no tokens. It drives the AI CLI you already use and pay for.
- Inspectable. Every run leaves a history and a log, so you can see what happened.
- Open source. The code is public under the MIT license.
04
Who builds it
LoopX is created and maintained by Encore Shao as an open-source project. Releases, source code and documentation live on GitHub, where issues and contributions are welcome.
Contact
Questions, corrections or requests about this page? Reach the maintainer here.