Introducing swarmctl

We're excited to announce swarmctl — a new way to provision ephemeral PostgreSQL databases on demand.

Why we built this

Modern AI agents and development workflows need databases that are fast to create, fully isolated, and disposable. Existing solutions either take too long to provision, share infrastructure between tenants, or require manual cleanup.

swarmctl solves this. Every database runs in its own Firecracker microVM with a dedicated kernel — true isolation, not just container boundaries. Databases boot in milliseconds, hibernate after 30 seconds of inactivity, and automatically destroy themselves when their TTL expires.

How it works

• Instant provisioning — spin up a full Postgres 17 instance in under a second
• Auto-hibernation — databases sleep after 30s of idle, wake in milliseconds on the next connection
• Self-destructing — set a TTL (default 1 hour) and the database cleans itself up
• Fully isolated — each database runs in its own Firecracker VM with up to 100GB storage
• Dual access — connect via standard PostgreSQL TCP or use the built-in REST API (PostgREST)

Get started

Install the CLI and create your first database in seconds:

curl -fsSL https://swarmctl.dev/install | sh
swarmctl auth login
swarmctl db create

Or sign in to the dashboard to manage databases from your browser.

What's next

This is just the beginning. We're working on snapshots, branching, team workspaces, and deeper integrations with AI agent frameworks. We'd love your feedback — reach out on GitHub or X.