Before you join the Beta
What you'll need to run LStack.
LStack is a local-first control plane you run on your own machine. Here's what it takes to run it today, who it's built for right now, and how Beta invitations work — so you know what you're signing up for.
What you'll need
Docker
LStack runs as a local Docker-Compose application today. You'll be comfortable cloning a repo and running a compose stack.
A model endpoint
A local or compatible OpenAI/Anthropic endpoint — LM Studio, Ollama, llama.cpp, or a hosted compatible API — for workers to run against.
A machine with some headroom
Workers and project runtimes run in containers on your hardware, so a workstation, Mac, or homelab server with capacity to spare works best.
Who it's for right now
Individual AI-assisted builders running local or compatible models who want more control than hosted coding agents give.
Solo founders & agent operators running several efforts at once who want throughput and a place to see status across them.
Homelab & workstation owners with spare local compute who want to put it to work as AI execution capacity.
This is early — what to expect
- Single-tenant and unauthenticated — built to run on your own machine, by you.
- Loopback-only; runtimes are exposed at localhost, with no LAN/WAN exposure.
- APIs and surfaces are still evolving — expect rough edges and frequent updates.
How Beta invitations work
We invite selected testers directly from the waitlist — there are no access codes to request or redeem. When it's your turn, we'll email you and add you to our private GitHub Beta repository at github.com/lstack-ai/beta, where you'll get the release, report issues, and join the discussion.
The Beta repository is private — access is granted to approved testers only.
By using LStack you accept the Terms of Service and acknowledge this is pre-release software.