Docs
Everything you need to deploy and run your app on Dockhold.
Dockhold takes a web app, builds it, and runs it at a live HTTPS URL. Two ways to deploy: connect a GitHub repo and push to your main branch, or run one command in a local folder with no repo. It works the same whether you wrote the code or your AI tool did.
Start here
QuickstartGo from a GitHub repo to a live URL. The short way.RecipesStep-by-step deploy guides for your stack: Vite, Next.js, FastAPI, and more.TroubleshootingFixes for the snags most apps hit on their first deploy.
Core concepts
How deploys workConnect a repo, push to your main branch, and your app goes live. Here is what happens in between.Connecting GitHubPublic repos just work. For private repos, pick the GitHub app or a personal token.Dockerfiles & dockhold.jsonWhat a Dockerfile is, when you need one, and copy-paste templates. Plus what dockhold.json does.Deploy from your computerNo repo? Ship a local folder to a live URL with one command.Environment variables & secretsSet plain config as variables. Keep API keys and passwords in the encrypted Vault.Databases & storageTurn on a managed database and read DATABASE_URL. Why your files reset, and what to do instead.App storageGive an app a folder that survives restarts and deploys. For a SQLite file, uploads, or a tool with its own data directory.Import a databaseAlready have data? Load it into your managed database once from a dump you upload.Connected servicesDeploy a frontend and an API together and Dockhold wires them up by URL for you.Runtime & limitsAlways-on apps, no cold starts. What each plan gives you for CPU and memory.Private appsLock an app behind an access token so only requests you allow can reach it.Custom domainsServe an app on your own domain with automatic HTTPS. Add one DNS record.Logs & monitoringWatch CPU, memory, and deploys live in your dashboard. No setup.
Deploying with an AI tool
An AI coding tool can deploy here on its own. Point it at our machine-readable guide at llms.txt (orllms-full.txt for everything in one file), or paste our agent rules into your project so the tool generates Dockhold-ready apps. You can also connect a tool to Dockhold's own deploy tools over MCP. SeeDeploy from your AI tool.