Skip to main content
Antideploy takes a project and gets it running in production. You don’t describe your app to it; it reads the code and works out what it is, what it depends on, how to build it, and how to start it.

Getting code in

GitHub

Connect a repository. Every push to the default branch redeploys.

Upload

Drag a folder or ZIP into the dashboard. No Git required.

API

Push a project directory over HTTP. What a coding agent uses.

What happens after that

1

Analysis

Antideploy reads your files and produces a spec: language, framework, build command, start command, port, and whether it needs a database. You can see exactly what it concluded, and the evidence for each conclusion, before anything is built.
2

Provisioning

If your code uses Postgres, a database is created and its connection string is injected as DATABASE_URL. A project that uses an S3 client gets a private bucket for its files, and your coding agent can also switch on sign-in for your app’s users and email for your app. Each one is wired into the environment for you. The platform overview lists every service.
3

Build

Your project is built into a container image. Static sites skip the buildpack step entirely and are layered onto a minimal server, which takes them from roughly three minutes to about seventy seconds.
4

Release

The container replaces the previous version and is given an HTTPS URL. Environment variables are injected at this point, and the deploy counts as live only once the new version answers.

What you never configure

  • Containers. No Dockerfile unless you want one; if you have one, it’s used.
  • Build and start commands. Inferred from your project, and shown to you.
  • Postgres. Created and wired up when your code needs it.
  • Files, sign-in and email. A private S3-compatible bucket, sign-up and sign-in for your users, and email sending are created for the app and wired into its environment.
  • TLS. Every app gets HTTPS.
  • Idle cost. Apps scale to zero and cost nothing while nobody is using them.

What Antideploy does not do yet

Being straight about this is more useful than a longer feature list.
  • Files written to disk are lost on restart or redeploy. Keep uploads in the private S3-compatible bucket Antideploy creates for the app.
  • Custom domains start on the Go plan. On the Free plan, an app is published only at its Antideploy address.
  • No long-running background workers. Each app is one HTTP container that sleeps when idle, so a worker or a polling loop will not stay alive. Antideploy tells you at analysis time when your project defines one.
  • Cron is supported. Up to three scheduled jobs per app, five minute minimum interval. Each one calls a path on your own app.
  • No staging or preview environments. One deployed version per project.
  • No horizontal scaling. Each app runs on a single instance that scales to zero when idle.
  • No Redis, MySQL or MongoDB. Supply your own, with its credentials as environment variables.
Antideploy raises these as hazards during analysis, before you deploy, rather than letting you discover them in production.

Operating a deployed app

Logs

Search and read your application’s output.

Deployments

Full history of every release, with the commit each GitHub deploy built.

Monitoring

Health checks, with email when an app goes down or recovers.

Environment variables

Configuration and secrets, encrypted at rest.

Start

Deploy something in five minutes

Connect a repository, upload a folder, or push over the API.