Deploy TypeScript apps on Prisma Compute
Push code, it runs. Your app runs as a long-lived TypeScript process on Bun, next to your database. A good fit for APIs and AI agents that hold state, stream responses, and keep connections open.
Built for the way TypeScript runs.
Push code, it runs
Long-lived runtime
TypeScript apps
Static assets included
Pairs with Prisma Postgres
Recovers on its own
Deploy from your repo
Run one command and Prisma Compute builds your app and brings it live with a URL.
@prisma/cli@latest app deployfrom the CLI, then connect a GitHub branch for push-to-deploy- No CI/CD pipeline to configure
- No deployment scripts, no dashboard workflows
- What's in your repo is what runs
Stateful execution
Your code runs as a long-lived process. Connections stay open. In-process caches persist across requests.
- Long-lived processes, not per-request functions
- Long-running HTTP requests and streaming responses
- Connections and in-process state persist across requests
Any TypeScript workload
Standard TypeScript on Bun, not a constrained edge runtime. No V8 isolate limits.
- Backend APIs and full-stack apps
- Long-running and streaming HTTP workloads
- AI agents, retrieval, and LLM orchestration
- Coming soon: background workers, cron, and WebSocket servers
Co-located database
Auto-wires Prisma Postgres and runs in the same environment.
- Low-latency access to the co-located database
- Built-in connection pooling for long-lived processes
- Pairs with a Prisma Postgres database in one step
- Works with any database, no lock-in
Does it work for me?
Deploying Next.js today?
Deploy with one command. Your assets are served by your app the way they are when self-hosting Next, with cache headers under control.
Running a Hono, Express, or Fastify API?
Long-lived processes are the default, so streaming and long-running requests work without workarounds. Connections stay open and streams run uninterrupted.
Building an AI agent?
Long-lived runtimes and in-process memory, on the same runtime that serves your API.
Need background jobs or workers?
Coming soon. Run them as long-lived processes alongside your API, in the same repo and runtime.
Bringing your own database?
Pair with Prisma Postgres for zero-config provisioning, or connect to anything else. Either way, no lock-in.
Tired of the serverless tax?
Tired of paying extra for expensive egress, image transformations, and third-party realtime services? With Prisma Compute, it all comes standard on a powerful Bun runtime.
One service for your app and its assets.
Starter apps
Next.js
A full-stack Next.js app paired with Prisma Postgres.
$ bunx create-prisma@latest --template nextTanStack Start
Modern full-stack TypeScript with Router & Query, paired with Prisma Postgres.
$ bunx create-prisma@latest --template tanstack-startHono API
Lightweight API backend. A good starting point for TypeScript APIs.
$ bunx create-prisma@latest --template honoHow it works
Why we built Prisma Compute
We wanted to ship TypeScript apps without stitching together separate build, hosting, and database tools. Push a repo, get a URL, and keep working in your codebase.