Build in a dorm,
Scale to billions.
Launchd gives you the pieces to get a real product online without spending weeks wiring infrastructure together.
Infra for teams that need to move fast.
Compute, Postgres, storage, auth, queues, and observability in one place. You should not need a whole platform team just to launch something real.
Run close to users without making deployment complicated.
Replicas, failover, and one connection string you can actually reason about.
No hidden analytics in the SDK. No weird surprises later.
Use your own S3 setup or ours. Either way, it stays simple.
JWT, OIDC, and magic links without a giant SDK taking over your app.
Logs, traces, and metrics where you need them.
Not just another AWS reseller.
We own the hardware and build the infrastructure ourselves. That gives us more control over performance, pricing, and how the platform actually works.
Infrastructure as code that still feels like code.
Declare the system once: functions, containers, Postgres, queues, storage, auth, realtime, and regions. Launchd turns that into the boring operational pieces without making you become the platform team.
$ launchd plan --env production ◆ graph compiled from Launchfile ├─ fn/api edge runtime · iad,lhr,syd ├─ pg/main primary iad · read replicas 2 ├─ queue/events durable · retry 5 · dlq on ├─ bucket/assets versioned · public cdn ├─ auth oidc · magic links · jwt └─ worker/billing container · cron 15m + 12 resources to create ~ 03 bindings to update = zero app rewrites
→ provisioning postgres → warming edge routes → binding secrets → replaying migrations ✓ stable in 3 regions
policy no public db policy queues require dlq policy secrets stay server-side pass deploy can proceed
Most backend platforms break the moment your project stops being a demo.
We kept running into the same problem over and over again.
The tools made for small teams were great right up until the product became real. You could launch quickly, but the second traffic showed up you were suddenly stitching together five different providers, rebuilding half your stack, rewriting auth, moving storage somewhere else, adding queues, adding containers, adding background jobs, and explaining to investors why your architecture diagram looked like a subway map.
Then on the other side you had the enterprise platforms. Powerful, stable, infinitely configurable, and completely unreasonable for a team trying to build something in a weekend. Before you could even ship a feature you were already dealing with networking policies, Kubernetes manifests, IAM rules, observability pipelines, container registries, and dashboards filled with words nobody on a startup team should have to think about at 2AM.
There was never a real middle ground.
Either you accepted a platform that worked for prototypes but fought you later, or you hired infrastructure engineers before you even knew whether people wanted the product.
Launchd exists because we wanted something that felt different. Something that let a single developer move fast without quietly building technical debt into every decision. A platform where adding real infrastructure was not treated like "graduating" from the simple product. A platform where functions, containers, Postgres, storage, queues, background jobs, auth, realtime systems, and edge compute all belonged together from the start instead of becoming emergency migrations six months later.
We are not trying to hide infrastructure behind fake abstractions. We are trying to make infrastructure feel proportional again.
If you want to deploy a single API, you should be able to do that in minutes. If your app suddenly ends up on the front page of Hacker News, you should not have to rebuild the company around surviving it.
No rewrites later.
The same platform that runs your first side project should still make sense when you have customers, databases in multiple regions, workers, queues, and real traffic.
No infrastructure cosplay.
You should not need a dedicated DevOps team just to launch an application that makes money.
No platform traps.
Standard tools. Predictable systems. Infrastructure that behaves the way developers already expect it to.
A small team trying to make infra feel less annoying.
We care about making the developer experience feel clear, fast, and sane.
Keeps the company pointed at the product we actually want to build.
Keeps the company moving and makes sure everything stays organized behind the scenes.
Leads platform engineering and keeps the developer experience simple.
We're not live yet.
Get on the list.
We're still finishing things up. Drop your name and email and we'll send access when it is ready.