Designed for hostile input.

Content-sanitization tools process untrusted text, URLs and archives. WordMark Remove treats those inputs as hostile by default.

Website scanner

• HTTP/HTTPS only.
• DNS resolution is checked and private/reserved IP ranges are blocked before every fetch and redirect.
• Credential-bearing URLs, localhost and local/internal hostnames are rejected.
• Redirects are manually followed and revalidated.
• Same-origin crawl only, robots.txt respected, page/body size and page-count caps applied.

ZIP processing

• Absolute paths, traversal paths and Windows drive paths are rejected.
• Entry count, input size and estimated expansion size are capped as fair-use abuse controls.
• Unsupported entries remain unchanged instead of being interpreted unsafely.

Accounts

• Passwords use scrypt with per-password random salts.
• Sessions are HMAC-signed, HttpOnly, SameSite=Lax and Secure in production.
• Sync payloads use AES-256-GCM encryption at rest.

Headers

The Next.js configuration sets HSTS, clickjacking protection, MIME sniffing protection, restrictive permissions policy and cross-origin isolation headers appropriate to this application.