RedSight scans your app the way an attacker would. Exposed databases. Leaked keys. Injectable endpoints. With proof.
Click through the interface. Expand findings. Toggle data masking. This is the real product.
01-recon — Fingerprinting, subdomain discovery, admin panel detection. Found: Supabase, React, Vercel deployment.
Not a checklist of best practices. Every finding includes the exact endpoint, the evidence we extracted, and how to fix it.
Service role key found in client-side JS bundle. Bypasses all RLS policies.
Sort parameter accepts arbitrary SQL. 5-second delay confirms execution.
Unauthenticated users can write to the root path. All app data is exposed.
API responds with Access-Control-Allow-Origin: * combined with credentials.
No agents to install. No config files. Just a URL and 5 minutes.
Paste the URL of your app. Works with any stack — Lovable, Base44, Bolt, v0, or custom builds on Supabase, Firebase, and beyond.
Reconnaissance, injection testing, bundle analysis, BaaS exploitation, auth attacks, SSL checks — everything runs simultaneously.
Every finding comes with the exact endpoint, evidence, severity score, and remediation steps. No false positives — we prove it or we don't report it.
AI code generators skip security steps. Attackers know this. Here's the chain we find — and break.
Most scanners check headers and call it done. RedSight actually attacks your app — safely — and shows you the proof.
Every finding includes the exact request, response, and evidence. If we report SQL injection, we show the extracted data.
Specialized scanners for Supabase, Firebase, Lovable, Base44, Bolt, and v0. We know where AI generators cut corners.
No agents, no SDKs, no YAML files. Paste your URL and get results in 5 minutes. Works from the outside, like a real attacker.
Every scan gets a 0-100 security score with letter grade. Track improvement over time as you fix issues.
Connect your repo and get pull requests with fixes generated automatically. Review and merge — done.
We don't just find vulnerabilities. We check if data has already been extracted through the holes we discover.
Every scan runs all 17 scanners. You only pay to see the full details.
Everything you need to know about RedSight.
RedSight is built specifically for AI-generated apps. We test for the exact patterns that Lovable, Bolt, v0, Base44, and similar platforms produce — like exposed Supabase keys, misconfigured Firebase rules, and missing auth middleware. Every finding includes actual evidence, not just theoretical risks.
Yes, completely safe. RedSight sends read-only HTTP requests — exactly what a browser does. We never write data, modify files, or run code on your server.
Yes. Before scanning, you verify ownership via DNS TXT record, a file upload, or an HTML meta tag. This prevents unauthorized scanning of sites you don't own.
Every free scan runs all 17 scanners and shows you a security grade, severity breakdown, and finding titles — so you know exactly what's wrong. But the details (exploit evidence, affected endpoints, remediation steps) are blurred. Unlock the full audit for $39 one-time, or subscribe to Pro ($49/mo) for unlimited audits and deep scans.
A complete security audit report for one domain: every vulnerability explained with evidence, step-by-step remediation instructions, attack chain analysis, and a downloadable HTML report you can share with your team. No subscription — pay once, keep the report forever.
Yes. If you subscribed to our previous Solo plan, your account is automatically mapped to Pro with the same (or better) limits. No action needed.
Quick: ~2 minutes. Standard: ~5 minutes. Deep: up to 10 minutes. 17 scanners run in parallel for maximum coverage.
Any web application accessible via URL. We have specialized scanners for Supabase, Firebase, WordPress, Shopify, Drupal, and Joomla — but the core scanners work with any stack.
Free scan. No credit card. No setup. Just paste your URL and see what's exposed.
Start Free Scan