Azure AIInstall

Summary

Analyze each input image with Azure Content Safety, checking for hate, violence, self-harm, and sexual content. Choose a publicly accessible image URL or a binary property from the input item, then set the severity threshold used to classify the result as safe or unsafe.

Inputs

• Authentication (required) — Choose Azure resource API key authentication or Microsoft Entra ID OAuth2.
• Image Source (required) — Choose URL or Binary File as the image source.
• Image URL (required) — When Image Source is URL, provide a publicly accessible image URL.
• Binary Property (required) — When Image Source is Binary File, name the input item's binary property containing the image. Defaults to data.
• Safety Threshold — Severity from 0 to 6 at or above which a category makes the result unsafe. Defaults to 2; lower values are stricter.

Output shape

One JSON item per input item, containing safe and categories fields.

safe is false if any returned category severity meets or exceeds the threshold. categories maps category names to severity values; Azure analyzes Hate, Violence, SelfHarm, and Sexual. The image is sent to Azure as a URL or base64-encoded binary content.

Links

• n8n community nodes installation guide