Fibe Blog

Most of your stacks are supposed to stay up. A web app, a worker, a dev server — you launch it, it serves, you keep it running. But a lot of the work you actually run against a project is the opposite shape: it should start, do one thing, and stop. Tests. A migration. A nightly backup. CI on a push. For that work, a long-running Playground is the wrong tool — it has no idea when it is finished, so it just sits there, holding resources, waiting for a URL that nobody is going to hit.

A Trick is Fibe's answer for the second shape. It is the same environment engine as a Playground — same Compose file, same Marquee, same logs and terminal — but Fibe treats it as one-shot: it runs the watched service to completion, records whether that service exited cleanly, tears the containers down, and keeps the result. This guide walks you through building one from scratch, with a real test-runner example, and explains the three things people get wrong the first time.

Every Playground on Fibe is born with an expiration date. Spin one up, walk away, and eight hours later it's gone — containers down, volumes reclaimed, the database row deleted. That's not a quota we enforce grudgingly; it's a default we're proud of. Idle environments are a tax everyone pays: they burn compute on a Marquee someone funds daily, pile up as ghost subdomains, and make "what's actually running here?" unanswerable. So we let them die.

But there's exactly one thing we will not do, no matter how expired an environment is: throw away your uncommitted work. If a Playground has dirty files when its clock runs out, the expiry sweep parks it rather than destroying it. This post is about that one rule — and how little code it takes to hold ephemerality and data safety at once.

The first time someone watches a Fibe environment come up, the question is always the same: where is this thing actually running? You clicked a button, a URL appeared, and thirty seconds later you had a full-stack app with its own subdomain and a valid TLS cert. No docker run, no DNS records, no certbot. It feels like magic, and magic is exactly the kind of thing engineers distrust.

So let's pull the curtain back. This post is the map of the whole platform — the nouns, how they fit together, and the literal sequence of events between "click Launch" and "your app is live." If you read one thing about how Fibe is built, read this; the rest of the blog drills into the corners.

Every coding agent on Fibe — a "Genie" — runs as a Docker Compose project on someone's Marquee. When a Player clicks "start," we want one thing to be unambiguously true: at most one live Genie chat per agent on that host. Not "usually one." Not "one, unless two browser tabs fire at the same millisecond." One.

The interesting part isn't the rule. It's that we enforce it in two completely independent places — the application layer reuses the existing chat, and the database has a unique index that physically cannot hold two — and that this redundancy is the point, not an oversight. Here's why, and what leaning on a real invariant buys you downstream.

You already have a docker-compose.yml. It runs on your laptop, it has a web service and a database, and you've typed docker compose up enough times to have it memorized. The promise of Fibe is that the same file — almost unchanged — becomes a launchable environment with a real HTTPS URL, logs, and a terminal, running on a host you control. No Kubernetes manifests, no rewrite, no new DSL to learn.

The trick is that a Fibe template is a Compose file. You don't convert it into something else; you add a few labels to the services that need them, and Fibe reads those to handle routing, TLS, and lifecycle for you. This post walks the minimal additions on a small, real example, explains what each one buys you, and ends with you clicking Launch and getting a URL.

Some people don't want us to conjure them a VM. They already have a box — a beefy machine under a desk, a hardened instance in a VPC their security team trusts, the staging server they've been paying for anyway — and they want their host to be the thing Fibe deploys onto.

So we let them. You hand Fibe an SSH endpoint — host, port, user, a private key — and that box becomes a Marquee: a first-class Docker host that runs Playgrounds, Tricks, and AgentChats just like one we provisioned ourselves. The interesting part is how little we had to add: a bring-your-own Marquee is the same Marquee, minus the part that conjures and tears down a VM.