Research Computing Environment

LaboratoryOSv0.4 — early access

Run your own AI infrastructure. Laboratory OS is the containerized Linux desktop environment purpose-built for self-hosted open-source AI.

Get Started

Overview

The ultimate workstation for open source AI.

From media pipelines to autonomous agents.

Media GenerationCUDA Accelerated

Generative Image, Video & Audio

Run the latest open-weight generative models on your own hardware. No usage limits, no usage fees, no data leaving your machine. Full creative control at unrestricted throughput.

Image GenerationText-to-VideoMusic GenerationTTS / STTComfyUI Workflows

Wan 2.2 · 848×480 · 4s

Video Generation with Wan 2.2 and ComfyUI

Build a text-to-video pipeline using Wan 2.2 inside a ComfyUI workflow — from model install to first render.

LLM HostingCUDA Accelerated

Self-Hosted Language Models

Serve state-of-the-art open-weight LLMs over a private OpenAI-compatible API. No per-token cost, no usage limits, predictable reliability. Your models, your infrastructure.

OpenAI-Compatible APIvLLM / Ollama / llama.cppNo Token CostsPrivate Endpoint

Summarize the key findings from this dataset.

Private LLM API with Qwen3 and vLLM

Deploy a production-grade OpenAI-compatible endpoint serving Qwen3-30B at full GPU throughput.

Agentic Computer UseNo GPU Required

Sandboxed Autonomous Agents

Give AI agents a fully isolated Linux desktop environment to work in safely. Interact with and supervise the agent through the live web desktop with file browser and VSCode.

Sandboxed OS EnvironmentWeb Desktop UISafe Privileged ExecutionLive Agent Supervision

{agent-id}.tunnels.laboratory.computer

Autonomous Coding Agent with OpenCode

Run OpenCode inside a sandboxed Laboratory desktop — full filesystem access, terminal, and live supervision from your browser.