Launching today
Alysis Code
Open-source AI coding agent
24 followers
Open-source AI coding agent
24 followers
Alysis is a powerful, open-source AI agent for desktop and terminal. Work with files, plan changes, write code, and review the results with your choice of models.
Free Options
Launch Team
Render Sandboxes: a computer for every agentRun untrusted code safely. Idle sandboxes cost nothing
Promoted
Hi Product Hunt! I'm Apollon, one of the founders and the lead engineer of a small applied AI lab in Athens.
We kept jumping between AI agents every time a new model or a new cool function/feature came out. So we built Alysis: one desktop app (Mac & Windows) where your work stays put and you just switch the model. It has 30+ built-in providers, plus local models.
It's not just for code. We use it every day for spreadsheets, docs, slides and research in its own browser. It also offers a built in computer use(still an early preview). Moreover it has a built in sandbox because safety is our top priority.
It's free with your own key, and plans start at $1/month if you don't want to deal with keys. The engine and CLI are open source.
It's early, so we now there are rough edges. The feedback that would help most is what breaks, and which everyday task you'd actually hand to an agent. Thanks for checking it out 🙏
I've been jumping between agents the same way — a new model drops and my setup feels stale overnight. The part I'd actually test first is whether my project rules and context survive a model switch cleanly, or if each model needs its own setup. If the plan step stays consistent across models, that's the bit that would keep me from switching again. How are you handling that — one shared instruction file, or per-model config?