Railway

Deploy bolt.diy | Browser AI App Builder, Pinned and in Production Mode

bolt.diy on Railway: pinned build, production mode, no keys demanded

Deploy bolt.diy | Browser AI App Builder, Pinned and in Production Mode

Deploy and Host bolt.diy on Railway

bolt.diy is an AI app builder that runs the whole project in your browser: you describe what you want, it writes the code, installs the packages and runs a dev server in a WebContainer, and you watch the preview update.

About Hosting bolt.diy

This template runs the official image pinned to a specific build, in production mode, and asks you for nothing before deploying.

Both of those are corrections. Upstream publishes only a moving latest tag and per-commit builds, so an unpinned template silently changes underneath you; this one pins the commit build. And the widely copied bolt.diy template on Railway sets NODE_ENV=development along with the Vite HMR variables from the local docker-compose profile - a dev server, with hot-reload pointed at localhost, behind a public proxy. That is not a configuration that works over the internet.

The other difference is what the deploy form asks for. The common template declares a dozen provider API keys with empty values, and Railway turns every empty value into a required field, so you have to supply keys for twelve different AI providers before the Deploy button will even light up. bolt.diy takes API keys in its own settings screen, per provider, whenever you actually want one. This template asks for nothing.

Common Use Cases

  • Prototyping a small web app by describing it, with the code and preview side by side
  • A scratchpad for trying a framework you do not have set up locally
  • A shared builder for a team that does not want to hand out local Node environments

Dependencies for bolt.diy Hosting

  • An API key for at least one model provider - Anthropic, OpenAI, Google, Groq, OpenRouter, Together, xAI, HuggingFace or a local Ollama. Add it in the app under Settings, not here.
  • Nothing else. Projects and chat history live in the browser, so the server keeps no state and needs no database or volume.

Deployment Dependencies

Implementation Details

VariableWhy
PORT, HOSTThe app listens on 5173 on all interfaces; PORT is what Railway routes the public domain to
NODE_ENVproduction - the image ships the built app, and dev mode would serve an HMR client that cannot reach you
VITE_LOG_LEVELinfo rather than debug, so the logs stay readable
DEFAULT_NUM_CTXDefault context window offered to local models

Because state lives in the browser, a redeploy costs you nothing and there is no volume to pay for. The flip side is worth knowing: clearing site data clears your projects, and two people opening the same URL do not share anything.

Why Deploy bolt.diy on Railway?

One service, one domain, no build to babysit. Pinned to a known build, so the version you tried is the version you get back after a redeploy.


Template Content

More templates in this category

View Template
Chat Chat
Chat Chat, your own unified chat and search to AI platform.

okisdev
113
View Template
stella
Self-host stella with web, API, Postgres, Redis, and object storage.

Jan Kubica
1
View Template
Hermes Agent | OpenClaw Alternative with Dashboard
Self-Hosted Hermes AI Agent for Telegram, Discord & Slack

codestorm
51