Where should I deploy an AI-built web app?
Pick the platform with first-class support for your framework and per-branch preview URLs. Previews are what let you review an agent's change as a running site instead of as a diff.
Surveyed 17 September 2026
Deploy an application
Open in Tools →- 01Vercel
Frontend cloud platform offering zero-config deploys, preview URLs, and edge functions.
Developer Tools - 02Netlify
Hosting for web projects, with CI/CD pipelines, serverless functions, and deployment to edge locations.
Developer Tools - 03Railway
Full-stack PaaS that deploys databases, backends, and cron jobs with zero configuration.
Developer Tools - 04Fly.io
Hosting platform that runs Docker-based apps on servers close to users, with persistent volumes and Postgres.
Developer Tools - 05Render
Cloud platform for deploying web services, static sites, cron jobs, and managed databases through Git-based workflows.
Developer Tools - 06Cloudflare Workers
V8 isolate serverless platform running in 300+ cities with low cold starts and KV, D1, and R2 storage.
Developer Tools - 07Google Cloud Run
Serverless platform on Google Cloud that runs stateless containers scaling from zero to thousands of instances.
Developer Tools - 08DigitalOcean
Cloud provider offering droplet VMs, managed databases, Kubernetes, and an app platform with a clean UX.
Developer Tools - 09Heroku
Platform-as-a-service supporting many languages with simple Git deploys and an add-on marketplace.
Developer Tools - 10Coolify
Self-hostable alternative to Heroku, Vercel, and Netlify for deploying from git and managing databases and services.
Developer Tools
A curated selection in editorial order. Use the fit and evidence to judge it for your task. Something missing?
What to look for
- 01Does every branch get its own URL automatically? This is the highest-leverage feature for agent-assisted work.
- 02Is the runtime a real one, or a constrained edge environment your dependencies will not run in?
- 03What does traffic cost at ten times current volume? Pricing cliffs are the usual unpleasant surprise.
Common questions
- Why do preview deployments matter so much with coding agents?
- Because they turn review into looking at the working thing. Reading a diff tells you what changed; opening the branch URL tells you whether it works.
- Is serverless the right default?
- For request-driven web apps, usually — you skip capacity planning entirely. Long-running work, websockets, and heavy background jobs still fit a persistent server better.
More in Ship and operate
- Interface with your agentsCLI harnesses, IDEs, control planes, desktop apps, multiplexers, and terminals for steering coding agents.
- Add a backendCombine databases, storage, APIs, and server-side functions.
- Add authenticationImplement accounts, sessions, identity providers, and authorization.
- Accept paymentsAdd subscriptions, checkout, billing, and payment infrastructure.
- Choose data infrastructureCompare databases, object storage, vector search, ORMs, and managed data services.
- Choose compute and GPUsFind cloud compute, serverless runtimes, and GPU providers for training and inference.
- Monitor product and usageCompare error monitoring, observability, product analytics, and web analytics.
- Automate deliveryBuild, test, preview, and release changes through CI/CD services.
- Add email and messagingSend transactional email, notifications, chat, and product messages.
- Protect agent credentialsCompare vaults, short-lived credentials, and egress proxies that keep secrets out of agent context.