Deploy, Convert & Scale

How to Deploy a Bolt.new App to Production (Domain, Env, SEO)

Bolt.new builds and runs your app in the browser using StackBlitz WebContainers, and you can publish straight from the editor to hosting connected to Bolt (Netlify has been the long-standing deployment integration). For a real production launch you also want the code in GitHub, secrets moved out of the frontend, a custom domain, a fix for SPA SEO and monitoring.

neooutput editorial team 4 min read Auf Deutsch lesen

Understand what Bolt.new actually produced

Before deploying, check which framework Bolt.new chose, because a Vite React SPA, an Astro site, a Next.js app and an Expo mobile project each deploy differently.

Open package.json and the config files, or simply ask Bolt to summarize the stack, build command and output directory. WebContainers run Node.js inside the browser tab, which is great for development but means some things that 'worked in Bolt' still need a real server environment in production, for example long-running processes, native binaries or background jobs. Identify those early so you are not surprised on launch day.

If Bolt produced a web app but you also want store apps, you do not necessarily have to start over in Expo. WebViewGold converts a deployed web app into native iOS and Android apps without code, with options for push notifications, deep links and in-app purchases. Be realistic about App Store review: under guideline 4.2, a wrapper needs real native value, not just the website in an app icon.

Publishing from Bolt with a custom domain

The quickest route is Bolt's built-in deploy, which publishes a production build and gives you a URL you can later map to your own domain.

If the deployment runs through Netlify, you can claim the site into your own Netlify account, which is worth doing for production: you then own the domain settings, environment variables, deploy history and rollbacks. Add the custom domain there, create the DNS records Netlify specifies (or delegate DNS to Netlify), and let it provision the HTTPS certificate. For single-page apps, make sure a redirect rule serves index.html for all routes.

  1. Run a production build inside Bolt and fix any build errors first.
  2. Deploy from Bolt and claim the site into your own hosting account.
  3. Add the custom domain and DNS records; wait for HTTPS.
  4. Add an SPA fallback redirect (for example a _redirects file with /* /index.html 200).
  5. Test deep links, login and payment flows on the real domain.

Get the code into GitHub

Production code should live in a Git repository with history, reviews and CI, not only in a browser project.

Use Bolt's GitHub integration or download the project and push it to a new repository. Connect that repository to your host so pushes to main deploy automatically and pull requests get preview URLs. From then on you can keep prompting in Bolt for features while engineers fix and review in the repo, but agree on one source of truth to avoid overwriting each other's changes.

Environment variables and secrets

Bolt projects often keep configuration in a .env file inside the project, and any variable exposed to the client bundle (such as VITE_ prefixed ones) is public.

Only public identifiers belong in client variables, for example a Supabase URL and anon key or a Stripe publishable key. Secret keys for payments, LLM APIs or admin access must be used only in server code: Supabase Edge Functions, Netlify Functions or a separate API. Set the production values in your host's environment settings, not in the repository, and rotate anything that was ever visible in a shared Bolt project or commit.

Database and auth: Supabase or Bolt's own backend

Whichever backend Bolt wired up, production means separate environments, strict access rules and backups.

With Supabase, enable Row Level Security on every table, write per-operation policies based on auth.uid(), set the production site URL and redirect URLs in Auth settings, and keep a separate project for staging. If you use Bolt's integrated database and hosting features, review the access rules and export options with the same rigor. In both cases, confirm backups exist and that you know how to restore them.

SEO for a Bolt.new site

If Bolt generated a client-side React app, search engines and social networks see a mostly empty HTML shell until JavaScript runs, so plan for prerendering or a static framework for public pages.

Ask Bolt to add per-page titles, descriptions and Open Graph tags, a sitemap.xml and robots.txt. For content that must rank, prompt Bolt to build marketing pages with a static-first framework such as Astro, or prerender key routes at build time. Verify in Google Search Console with the URL inspection tool that rendered HTML contains your content.

Monitoring after launch

Add error tracking, uptime monitoring and analytics before you announce the launch.

Install Sentry or a comparable tool with source maps, configure an uptime check on the domain and on any API or function endpoint, and watch your host's function logs and Supabase logs during the first days. Set alerts for error spikes and failed payments. A short incident checklist (who checks what, how to roll back the last deploy) turns a stressful outage into a routine fix.

Frequently asked questions

Can I use my own domain with Bolt.new?

Yes. Deploy from Bolt, then attach the domain in the hosting settings (for example after claiming the site into your Netlify account) and create the DNS records shown there.

Why does my Bolt.new app work in the editor but fail in production?

Common causes are missing environment variables on the host, an SPA without a fallback redirect, or code that relies on behavior specific to the in-browser WebContainer environment. Check the build and function logs first.

Is it safe to put API keys in a Bolt.new .env file?

Only public keys. Anything exposed to the client bundle is visible to users. Secret keys must be used exclusively in server-side functions with values set in the host's environment settings.

All insights