Get started
The cloud your agents
can operate.
LiveLLM is an AI-first cloud. Container apps, VMs, headless browsers and autonomous agents all live in a workspace, and every one is created, changed and torn down through the same REST API your engineers use — so your agents can drive it too.
Send your AI agent to LiveLLMdocs.live-llm.com/SKILL.md
⚡
Quickstart
From API key to a running workload in five minutes.
◇
API reference
Every endpoint, with copy-paste samples in cURL, TypeScript and Python.
✦
Agents & Skills
The agents that ship with every workspace — browsers, scaffolding, review.
🤖
For AI agents
A single SKILL.md your agent can read and act on.
What LiveLLM handles
You bring a spec; the platform brings the plumbing. Every resource is real and inspectable — nothing magical you can't reach.
☸
Container apps
Ship a prebuilt image or point at a repo in your workspace org — it builds from source (no Dockerfile) and rolls on every push.
🖥
VMs & browsers
Ubuntu, Ubuntu desktop and Windows VMs plus headless Chromium — each workload gets its own per-port HTTPS hostname.
🗄
A Git org per workspace
Every workspace comes with its own Git org: repos, mirrors and build status, all readable through the API.
✦
A fleet of agents
Browser agents with reviewable trajectories, an app scaffolder and code review — powered by the AI providers you connect.
Base URLs
The API lives on this host; your console is a separate app. Every exposed workload port gets its own dash-joined hostname under the shared wildcard certificate.
API
https://api.live-llm.com
Console
https://cloud.live-llm.com
Workload port
https://<id>-<port>-<workspace>.cloud.live-llm.com
e.g. a container app web on port 80 in workspace acme is served at https://web-80-acme.cloud.live-llm.com