HIVE_SECRET_KEY

From the Playground API Keys panel on thehive.ai — Get key ↗

SYSTEM_STATUS: AWAITING_SPECIMEN

⬡ DROP IMAGE HERE
OR CLICK TO SELECT

Synthetic Probability

--%

Origin Classification

UNDEFINED

HUD_SOURCE: — // SCAN_COMPLETE // ENGINE: HIVE_AI_v3

// Detection

Hive AI v3 synthetic-origin model. Score > 50% triggers MUTANT STATE. HUD boxes follow real anomaly coordinates.

// ELA Fallback

If the API returns no coordinates, a local Error Level Analysis scan identifies the three highest pixel-noise clusters.

// Privacy

Your Hive keys live in localStorage only. Images go direct to Hive AI — nothing passes through Mutant servers.

// FAQ

How do I verify if an image is AI-generated?

Drag and drop your image into the containment field above. Mutant Verify uses Hive AI V3 to return a synthetic probability score. Above 50% = likely AI-generated.

What is a C2PA Do Not Train assertion?

C2PA is an open standard (used by Adobe, BBC, Microsoft) for embedding cryptographic provenance into image files. A Do Not Train assertion is a declaration baked into the file structure — not just metadata — signalling the image should not be used to train AI models.

How do I add a Do Not Train assertion to my images?

Install the Mutantwork Power Pack for ComfyUI. Search comfyui-mutantwork in ComfyUI Manager. The Mutant Content Credentials node embeds a C2PA Do Not Train assertion into your PNG output. Free. MIT. Offline.

Can C2PA assertions be stripped from images?

Unlike EXIF/XMP metadata which strips easily, C2PA manifests are tamper-evident — altering the file breaks the cryptographic signature. Platforms supporting C2PA (Adobe Firefly, Getty, others) can read and respect these assertions automatically.

Is Mutant Verify free?

Yes. Completely free. No account needed. C2PA detection runs client-side. AI image detection uses Hive AI V3 — bring your own free API key in the ⚙ API KEYS panel.