YoubuiltitwithAI.Find the security gaps. Fix them.
Paste your site link. See what's risky, what to fix first, and get copy-paste instructions for Claude, Cursor, or any coding AI to fix it.
- No account to start
- We only look — never hack your site
- Email when it's ready
- leaked API keys in JS
- security headers & HTTPS
- .env, .git & backups
- Supabase / Firebase rules
- cookies & CORS
- vulnerable libraries
Findings
0 found · 0 locked
See it in action
Paste a link. Get the fix.
Paste your URL, watch the passive scan run, read the plain-English report, then hand one prompt to your AI to fix everything.
FixPrompt is a security scanner for apps built with AI tools like Lovable, Bolt, Cursor and v0. It checks what any visitor can see on your live site, such as exposed API keys, missing security headers and risky cookies, then gives you a plain-English report and a prompt your AI can use to fix it.
Learn more in the security guides or grab ready-made security prompts and rules.
- Safety checks on every scan
- 40+
- Free scan, no card
- 1
- Lowest price per full scan
- $1.95
- Expiry on credits. Ever.
- 0
How it works
AI wrote the code. Nobody checked if it's safe to ship.
FixPrompt does the safety check your build skipped. You can close the page — we'll email you when the report is ready.
Paste your URL
Any public http(s) app. Localhost, private IPs, and cloud metadata endpoints are blocked.
We look, we don't touch
We open your site the same way a visitor would — like a careful guest, not a hacker. We never change anything or run attacks.
Hand the fix to your AI
Copy one prompt into Claude, Cursor, Copilot, or any coding AI. It fixes issues in priority order and tells you how to double-check.
The report
See the risk. Unlock the fix.
Your free scan shows how many serious problems you have and what they're called. Unlock to see exactly what's wrong and how to fix it.
- Low findings open free after signup
- Medium → critical open with one credit, which also covers your re-scan after fixing
- One prompt fixes everything — paste it into any AI coding agent
Missing X-Content-Type-Options
Add X-Content-Type-Options: nosniff at the edge or in middleware to stop MIME sniffing.
› Fix prompt: add nosniff to every response in middleware.ts…
Client bundle references a privileged key name
Content-Security-Policy not set
Compare
Cheaper than a pentest. Faster than a checklist.
Where FixPrompt fits next to the other ways to answer “is my app leaking?”
Price
Time to a report
Knows AI-built stacks
Supabase / Firebase hosts, key names in client bundles
Fix prompt per finding
One prompt for Claude Code, Codex, Cursor, or any LLM
Ranked low → critical
Passive, no exploit payloads
No subscription
Deep manual logic testing
We're honest: for that, hire a human
Pricing
Pay per scan. Keep the credits forever.
Your first scan is free. After that, packs make each full, unlocked scan cheaper than a coffee.
Tip: scan once, fix, then scan again to prove it's gone.
Best fit · Builder · $19
$2.38per full scan
21% less per scan than the Starter pack
Starter
$9
3 full scans · $3.00/scan
One app: scan, fix, re-scan
- Every severity unlocked, low → critical
- AI fix prompt per finding + one fix-everything prompt
- Re-scan after fixing to verify
- Credits never expire
Builder
Save 21%$19
8 full scans · $2.38/scan
A couple of apps, or several rounds of fixing
- Every severity unlocked, low → critical
- AI fix prompt per finding + one fix-everything prompt
- Re-scan after fixing to verify
- Credits never expire
Studio
Save 35%$39
20 full scans · $1.95/scan
Agencies and side-project collectors
- Every severity unlocked, low → critical
- AI fix prompt per finding + one fix-everything prompt
- Re-scan after fixing to verify
- Credits never expire
Already ran your free scan? Unlock that report with 1 credit, straight from the report page.
One-time payments · no subscription · secure checkout
FAQ
Straight answers
Scope, the free scan, and exactly what you're paying for.
What is FixPrompt?
FixPrompt is a security scanner for apps built with AI coding tools such as Lovable, Bolt, Cursor and v0. Paste your URL and it passively checks what any visitor can see, such as exposed keys, missing security headers and risky cookies, then writes a plain-English report with a prompt your AI can use to fix each issue.
Is my Lovable, Bolt or Cursor app secure?
There is no way to know without checking. AI tools write working code, not necessarily safe code, and common gaps are secret keys in the browser, database tables without access rules and missing security headers. A FixPrompt scan checks the public side of your app in minutes, and your first scan is free.
What does the free scan include?
One free scan per account. Low-severity findings are unlocked, with the explanation and a fix prompt. Medium and above stay locked until you use a credit to unlock that report.
Is this a penetration test?
No. FixPrompt passively reads what any visitor's browser receives — pages, JavaScript bundles, headers, cookies, common public paths — plus public DNS records. It does not send exploit payloads or try credentials, and secret values it spots are never stored.
Which URLs can I scan?
Public http(s) URLs for apps you are authorized to assess. Localhost, private networks, and cloud metadata endpoints are blocked.
Do credits expire?
No. Packs are one-time purchases and credits never expire. There is no subscription and no automatic renewal.
Which AI can use the fix prompt?
Any of them. The report gives you one fix-everything prompt (plus one per finding) written for any coding agent with access to your repo: Claude Code, Codex, Cursor, GitHub Copilot, Windsurf, Gemini CLI, Cline, Aider — or a chat model via OpenRouter, DeepSeek, or GLM. It knows your stack, says where to change what, and how to verify each fix.
What does a credit buy, and which pack should I pick?
One credit is one fully unlocked scan, or the unlock of a report you already ran. Most people scan, fix and scan again, two to four scans per app, so the 8-scan pack is the usual pick; the 3-scan pack covers one app, and the 20-scan pack is for agencies or several apps. The bigger the pack, the cheaper each scan.
What happens right after I pay?
Checkout is secure and one-time. Your credits land in your balance as soon as the payment is confirmed (usually seconds). Spend one to unlock a report you already ran, or one per new scan.
Security guides
Built with Lovable, Bolt or Cursor?
Platform checklists with the mistakes AI tools make most, and a prompt to fix them.
- Lovable security
- Bolt security
- Supabase security
- Cursor security
- Next.js security
- Claude Code security
- v0 security
Blog
Security tips, no jargon
Short reads you can act on today.
- Supabase service_role key leaked? How to rotate it and recover
- How to check that Row Level Security is enabled on every Supabase table
- Cursor hooks for security: block risky shell commands and MCP calls
- A 10-minute security checklist before you launch a vibe-coded app
- How to check if your API keys are exposed in your JavaScript bundle
- Supabase anon key vs service_role key: which one can be public?
- 6 security headers every AI-built app should send
Find it before someone else does.
Paste a public URL. No account needed to start. Your first report is free.