Push git. You get https://name.usectl.com.
We run the metal.

Create a machine
My favorite project
Command

>usectl Pipeline
https://your-app.com
Successfully Deployed
Runtime: 1.4s • Hetzner DC

  1. 01 Create a machine.

    Postgres, Redis, S3, cron, domains. We run the metal.

  2. 02 Connect a repo.

    GitHub or the CLI.

  3. 03 Push.

    You get https://name.usectl.com.

$ usectl login
$ usectl machines create --repo … --postgres --redis
https://name.usectl.com

The CLI is the product.

Every capability of usectl is a single-binary CLI command. Structured --json output so your tooling can pipe it. Clear error messages so your AI agent can self-correct. Install once and every AI coding agent, whether Claude Code, Cursor, Windsurf, or ChatGPT, can deploy, read logs, and manage infrastructure just by running commands. No special plugin. No MCP connector required. Your agent already knows how to use a terminal.

cli ~/your-saas · usectl-cli v2.0.4
$ usectl auth login
✓ Authenticated as [email protected]
$ usectl machines create \
--repo github.com/you/your-saas \
--postgres --redis
✓ PostgreSQL ready. $DATABASE_URL injected
✓ Redis ready. $REDIS_URL injected
$ usectl deploy
Building with Kaniko...
✓ Build complete (28s)
✓ TLS certificate issued
✓ Live at https://your-saas.usectl.app
$ usectl logs --follow
[app] listening on :3000
[app] GET /api/health 200 12ms
[worker] processing queue...▊
WORKS WITH ANY AGENT Claude Claude ChatGPT ChatGPT Cursor Cursor Windsurf Windsurf
CLI docs

Prefer MCP? The built-in MCP server exposes the same commands as tool calls for tighter integration with Claude Code and Cursor. See the MCP docs →

On our Hetzner.

Postgres, Redis, S3, cron, domains. You push git.

  1. Git push. You get https://name.usectl.com.

  2. Postgres and Redis. Strings in the env.

  3. S3. Cron. Domains and TLS.

  4. CLI. Same commands as the how-to.

Deploy your first project

Your full stack live in under five minutes.

Serious infrastructure.
Zero configuration.

Every project runs in its own Kubernetes namespace. Isolated compute, isolated databases, managed storage. Production-grade infrastructure you don't have to configure.

We swapped MinIO for SeaweedFS to scale object storage better per tenant. S3 API is the same, so your AWS SDK code keeps working.