Spexop UI is an early, free, local-first, open-source UI/UX desktop companion in development. I am sharing the direction and the current evidence separately so readers can tell what exists today from what we still intend to build.
The intended workflow
A developer keeps their existing strong coding agent in charge of planning, design decisions, risk, and oversight. User-configured free or low-cost LLM API workers are intended to take on bounded routine UI tasks. Spexop coordinates a small milestone, gates, and tasks in the background. It should verify each result with deterministic checks and a real browser when the task needs one, then show plain progress, blocked, and needs-user-action states.
The goal is to make small UI changes easier to delegate and inspect without losing the developer’s control of the project.
What we have actually checked
Current evidence is a private local coordination prototype that uses test doubles, plus a loopback browser status view. The prototype has technical tests and a real-browser observation of the local status flow. Those results support only that narrow prototype.
There is no packaged native desktop app, no verified live LLM API worker, and no external product acceptance yet. The Groq access path remains blocked by a provider-side 403 response. The public-source review is still open, so the current monorepo is not being presented as a public release. Earlier Theme, Foundation, Capture, and icon work are technical assets, not evidence that this new companion is ready to install.
Where to follow and contribute
The current product direction and status are on spexop.com. You can also follow the Spexop UI LinkedIn page.
I would value feedback on one question: what is the smallest UI task you would trust a bounded worker to do if a coding agent planned it and independent checks reported the result?
Top comments (0)