Duvaxa

Security & Compliance · Security check for AI-built apps

LaunchLock

Built it with AI? Make sure it isn't leaking before you launch.

Finds the secret keys, public files and open database tables that apps built with Lovable, Bolt, v0, Replit and Cursor ship by accident, before a stranger or a customer's security team does.

Founders who built their MVP with Lovable, Bolt, v0 or ReplitAgencies shipping AI-built apps for clientsStartups about to face their first customer security review

Try the free version: AI-built app leak check → · Security headers check →

LaunchLock scans an AI-built app before launch and finds a secret service key in the JavaScript, a downloadable .env file and database tables anyone can read or edit. Sample data.
FindsLeaked keys and public files
TestsSupabase row-level security
Works withLovable, Bolt, v0, Replit, Cursor
SafeRead-only, keys masked

The problem

AI app builders get a working product live in a day, and the quickest way to make something work is often to put a key in the front end or leave row-level security off. Everything in the front end is downloaded by every visitor. A Supabase service-role key or an open table means anyone can read every user's data, and an OpenAI or Stripe key means anyone can spend your money.

Example output

launch-lock · example
$ launch-lock rls https://<your-project>.supabase.co --key $SUPABASE_ANON_KEY
grade F (key role: anon)
[CRITICAL] profiles: Anyone with your public key can read 1240 row(s) sensitive columns: email
[HIGH] posts: Anyone with your public key can read 12 row(s)
[ok] orders: No rows visible (row-level security filtered them, or the table is empty)
Fix: alter table profiles enable row level security; then add a policy using (auth.uid() = user_id)
Run against LaunchLock's test project, which has the same mistakes as many AI-built apps.

What it does

01

Scans the live app the way a browser sees it, including lazy-loaded pages: Supabase service-role, OpenAI, Anthropic, OpenRouter, Groq, Stripe, AWS, Google service account, Slack, Twilio and more

02

Finds public .env files, exposed .git folders and source maps

03

Tests your Supabase row-level security with your own public key: which tables strangers can read, how many rows, and which columns look personal

04

Tests your Firebase rules the same way: Firestore collections, the Realtime Database and Storage

05

One branded report for each client: your logo and name, every finding with its fix, what was and wasn't checked, and a sign-off box, ready to print to PDF (white-label on the agency plan)

06

Flags public storage buckets and database functions callable by anyone

07

Read-only: never writes to your database, and refuses service-role keys

08

Masks every key it finds in its output

09

GitHub Action: fail the deploy when the grade drops

How it works

  1. Run the free check on your app's address
  2. Run LaunchLock with your Supabase public key
  3. Fix what it finds, with the exact SQL or setting to change

Works with

SupabaseFirebaseLovableBoltv0ReplitCursorVercelNetlifyGitHub Actions

Your data

  • The app scan reads only public pages, like a browser.
  • The RLS check runs on your computer with your own public key and only reads.
  • Only test apps you own or have permission to test.

Who it's for

Founders who built their MVP with Lovable, Bolt, v0 or ReplitAgencies shipping AI-built apps for clientsStartups about to face their first customer security review

Questions

Isn't the Supabase anon key meant to be public?

Yes. That's why row-level security matters: it's the only thing between that key and your data. LaunchLock tests it table by table.

Do you need my service-role key?

No, and LaunchLock refuses it. The point is to see what a stranger with your public key can do.

Does it replace a penetration test?

No. It catches the most common, most serious mistakes in minutes. Logins, payments and write access are covered in our pre-launch audit.

Guides

Read more about this problem

Related

Often used with LaunchLock

Security & Compliance

MaskWall

AI data firewall

Use AI. Keep your secrets. Stops staff from accidentally sending passwords, API keys or customer data to ChatGPT, Claude and other AI tools. Sensitive text is hidden before it leaves and put back in the answer, so people can still use AI safely.

  • Who it's forBanks, insurers and fintechs · Healthcare, legal and IT services firms holding client data
  • What you getData-protection gateway · Claude Code and browser protection · Audit report for compliance
Free to start · from $8/user/monthView product →
Security & Compliance

ExpiryWatch

Certificate and domain expiry monitor

Nothing expires by surprise. Keeps a list of every website certificate and domain you own, checks them daily, and warns you in Slack or Teams before anything expires and takes your site offline.

  • Who it's forAgencies and MSPs with many client domains · Companies with many brands or country sites
  • What you getCertificate and domain inventory · Daily expiry alerts · DNS and email security fixes
Free to start · from $29/monthView product →
Security & Compliance

PrivacyPilot

DPDP privacy compliance

Privacy compliance, made practical. Checks your website against India's new privacy law (DPDP Act) and gives you the privacy notice, data list and cookie consent banner you need, with GDPR and other countries' rules checked too.

  • Who it's forClinics, schools and coaching centres · Web agencies and CA firms
  • What you getDPDP gap report · Privacy notice draft · Data inventory
Free to start · from $49/monthView product →

See LaunchLock on your own data

A 30-minute call. We run the product on your export, repository or document and walk you through the results.

Book a demo