Guides / Bolt
Is Bolt secure? A pre-launch security checklist
Bolt can take you from prompt to deployed app in an afternoon. Before you share it, check these common gaps in Bolt-built apps, from leaked keys to open databases.
Bolt makes it possible to build and deploy a full app without writing code yourself. That's exactly why a security check matters: you may not have read the code that decides who can see your data. The platform being trustworthy doesn't make every app built on it safe by default.
Use this checklist before you send your Bolt app to real users.
1. Find out where your data lives
Ask Bolt (or check your project) which database and auth the app uses. Most Bolt apps use Supabase or a similar backend that the browser talks to directly. The rules that protect that data are your responsibility.
- Supabase: make sure Row Level Security is on for every table. See the Supabase RLS checklist.
- Firebase: make sure your Security Rules aren't still in "test mode". See the Firebase rules checklist.
2. Look for secret keys in the frontend
Bolt projects are usually Vite apps. Any environment variable whose name starts with VITE_ is built into the JavaScript that every visitor downloads. That's fine for public values like a Supabase URL or anon key, and dangerous for secrets.
Secrets that must never have a VITE_ prefix or appear in frontend code:
- OpenAI, Anthropic, Groq or other AI keys
- Stripe secret keys (
sk_live_…) - The Supabase service_role /
sb_secret_key - Email service keys (Resend, SendGrid, Mailgun)
Move any call that needs a secret into a server function and read the key from a server-side environment variable. More detail: leaked API keys in your JavaScript.
3. Check sign-up, login and password reset
Try the flows yourself, logged out, in a private window:
- Can you open pages that should need a login by typing their URL?
- Does "forgot password" leak whether an email has an account?
- Can a normal user reach admin pages or other users' data by changing an ID in the URL?
If any of these work, the protection is only in the interface. Ask Bolt to enforce it in database rules or server code.
4. Don't publish what you didn't mean to
- Turn off production source maps, which publish your original code.
- Make sure files like
.envaren't copied into the public folder. - Remove test pages, seed data and debug routes before launch.
5. Add security headers
Your host serves the pages, so that's where headers go (Netlify, Vercel or Cloudflare usually). Start with Strict-Transport-Security, X-Content-Type-Options and a Content-Security-Policy. See how to add headers on each host.
6. Check speed on a phone
Not security, but it decides whether visitors stay. AI-built apps often ship large bundles and uncompressed images. Aim for the main content to appear in under 2.5 seconds on mobile.
7. Re-scan after changes
Every big prompt can change code you checked earlier. Make re-checking part of every release.
Check it in one minute
Scanora scans your live Bolt app for leaked keys, exposed files and source maps, missing headers, SSL problems, vulnerable libraries and speed, gives you an A–F grade, and shows the fix for each problem.
Check your site in one minute
Scanora runs these checks for you: leaked keys, open files, headers, SSL, speed and AI-search readiness. Free, no signup.