AI malware scanner for WordPress

What it does

Signature databases only catch malware someone has already catalogued. An AI model reads the actual code and gives a verdict on what it does, which is how obfuscated, novel or targeted backdoors get caught.

Unlike scanners that route your files through the vendor's cloud, this one connects directly to the AI provider you pick, with your own API key. There is no middleman seeing your files.

How it works

Pick an engine and model (Google's Gemini Flash is the default), add your key, and scan on demand or on a schedule from hourly to weekly. Large files are sent in chunks, and a single failed chunk skips rather than aborting the whole scan. Flagged files land in the quarantine workflow; known-good files go on the whitelist so they are never flagged again.

A separate heuristic scanner covers non-AI pattern checks, and continuous file monitoring flags changed files on its hourly scan: the natural trigger for an AI verdict.

• Google, OpenAI, OpenRouter or Requesty, with per-engine model choice
• On-demand and scheduled scans (hourly to weekly)
• Chunked uploads with per-chunk failure tolerance
• Quarantine and whitelist workflows
• Optional email notification of results
• Companion heuristic scanner for non-AI checks

Costs and privacy, plainly

The AI provider bills you for API usage under your own agreement with them, typically pennies per scan at current model pricing, but that is between you and your provider. Scanned file contents are transmitted to that provider; choose one whose data handling terms you accept.