DEVELOPER DOCUMENTATION

One clean API.
Every compute surface.

Everything you need to launch web apps, managed PostgreSQL 17, AI agent sandboxes, Ubuntu cloud desktops, and GitHub Actions CI runners from your terminal, codebase, or MCP-enabled AI assistant.

CLI REFERENCE

Essential uncage commands.

Walkthrough with real output →
uncage status
Verify connectivity to the uncage.sh control plane API.
uncage projects create --name <name> --region <region>
Create an isolated project namespace with private internal DNS.
uncage workloads create --project <id> --name <name> --surface <surface> --image <oci> --vcpu 2 --mem 4096
Boot a web app, database, AI sandbox, cloud desktop, or CI runner.
uncage workloads exec <id> -- <cmd...>
Run a command inside a running workload and stream stdout, stderr, and exit code.
uncage workloads fork <id> --name <branch>
Create a sub-200ms copy-on-write memory + disk clone of a running workload.
uncage workloads pause <id>
Snapshot live RAM + disk state and drop compute billing to $0.00/hr.
uncage workloads resume <id>
Wake a sleeping workload with processes and filesystem preserved.
uncage workloads list --project <id>
Inspect workload status, private DNS, public HTTPS URL, and specs.
REST API OVERVIEW

Direct HTTP access at https://api.uncage.sh

Authenticate every request with Authorization: Bearer ucg_live_... and JSON request/response bodies.

POST /v1/projects
Create a project in us-east, eu-central, us-west, or ap-south.
POST /v1/workloads
Provision any compute surface (app, database, sandbox, desktop, ci_runner).
GET /v1/workloads
List workloads for your organization, optionally filtered by project_id.
POST /v1/workloads/:id/exec
Execute commands synchronously inside a running workload with timeout_sec.
POST /v1/workloads/:id/fork
Live copy-on-write clone of memory and NVMe disk state.
POST /v1/workloads/:id/pause
Pause workload compute to $0.00/hr while preserving state.
POST /v1/workloads/:id/resume
Hydrate and resume a sleeping workload in sub-second time.
DELETE /v1/workloads/:id
Terminate a workload and release its compute allocation.
GLOBAL REGION CODES
us-east
Virginia, US
Typical edge RTT ~14ms
eu-central
Frankfurt, DE
Typical edge RTT ~19ms
us-west
Oregon, US
Typical edge RTT ~28ms
ap-south
Singapore, SG
Typical edge RTT ~42ms
DOCS FAQ

Common developer questions.

01 Which regions are available via the CLI, SDKs, and MCP server?
All four global regions—us-east (Virginia), eu-central (Frankfurt), us-west (Oregon), and ap-south (Singapore)—are available through the exact same region parameter across every surface.
02 Do all five compute surfaces share one API and billing account?
Yes. Apps, Databases, AI Sandboxes, Cloud Desktops, and CI Runners all use the unified /v1/workloads API, a single Bearer API key, and active-usage billing ($0.036/vCPU-hr, $0.0054/GB-hr, $0.00/hr idle).
03 Is there a free sandbox or trial to test the CLI and SDKs?
Yes. Create an account at /register to get instant access to the dashboard, API keys, and CLI. Outbound open internet egress unlocks automatically once your account has a $5 verified balance.
READY TO BUILD?

Grab an API key and launch your first microVM in 60 seconds.

Start with $10 in free credit. Zero idle billing across every surface.

✓ $10 starter credit on signup
✓ Strictly $0.00/hr compute when sleeping
✓ Sub-second boot across 4 global regions