Early access · Supabase change lands Oct 30, 2026

Your vibe-coded app works.
Is it safe to ship?

ProdCanary reviews your own Lovable, Bolt or Cursor project for leaked keys, weak Supabase policies, open functions and storage, then hands you a plain-English report and fix prompts you can paste straight into your AI editor.

Free scan: your top 3 risks in seconds

Upload a .zip of your own project (Lovable, Bolt, Cursor, Next.js, Supabase, Firebase). We review the code statically. We never contact your live site, and we never run your code.

Tip: leave out node_modules, .git, build output and images. Your upload is processed in memory or a temp folder, deleted right after the scan, and never stored. Secrets in results are masked.

RLS is on. That's not the same as secure.

We reviewed 30 public Lovable and Bolt apps built on Supabase. Here's what we found.

33%had at least one critical or high-severity issue
40%had privileged database functions any visitor could call to change data
13%had write policies open to everyone
7%allowed public upload or delete in storage
0%lacked RLS entirely. The danger is in the policies and functions around it.
1 appshipped a live service_role key in browser code

Small sample: n=30 public GitHub repos, reviewed from source. Treat these as directional, not as industry-wide rates.

Heads up: Supabase changes on Oct 30, 2026

New tables in the public schema will no longer be exposed to the Data and GraphQL APIs automatically. When things break, the quick fix is to grant anon broad access, and that can quietly open your data to anyone. ProdCanary flags overly broad grants and policies before you ship them.

Read the Supabase changelog →

How it works

1

Upload or connect your project

Upload a zip of your project, with GitHub connect coming soon. Read-only, and only your own code.

2

We review code and config

Client bundles, Supabase migrations and policies, functions, storage, Firebase rules and API routes.

3

Fix with copy-paste prompts

A plain-English report, ranked by severity, with a fix prompt for Cursor or Lovable on every finding.

A sample report

Generated from a demo project with deliberately planted flaws. No real app was used.

ProdCanary sample report showing critical and high findings with fix prompts

Pricing

Start free. Pay only if you want every finding and the fix prompts.

Test mode: checkout runs in Stripe's sandbox while we finish setup, so no real charges are made. Use card 4242 4242 4242 4242.

Free scan

$0

  • Top 3 most severe findings
  • One-line fix for each
  • Count of everything else
Run a free scan

Full report

$29 one-time

  • Every finding, explained in plain English
  • Copy-paste AI fix prompt for each issue
  • Upload the same zip after paying
Get the full report – $29

Monitor

$19/mo

  • Re-scan anytime
  • Alerts on new issues
  • Includes the full report
Start Monitor – $19/mo

FAQ

What happens to my code?

ProdCanary reads your code read-only, never modifies it, and deletes it after the scan finishes.

Is this a penetration test?

No. It's an automated review of your source and config. It catches common, high-impact mistakes but doesn't replace a professional pentest or audit.

Do you scan live websites?

No. ProdCanary only reviews projects you upload or connect yourself. It never probes anyone's live site.

Which stacks are supported?

Lovable, Bolt and Cursor-built apps, with a focus on Supabase and Firebase backends and JS/TS frontends.

Will there be false alarms?

Some. We test against real projects to keep them low, and every finding explains why it was flagged so you can judge for yourself.

Ship without the "oh no" moment.

Join the waitlist for early access and launch pricing.

No spam. Unsubscribe anytime.