How AI Media works with the AI module

AI Media adds no AI rules of its own. Every request goes through the AI module, so the site's own guardrails, moderation, prompts, context, logging and observability apply to it. AI Media says who it is, and reacts where those systems decide.

Piece of the AI module	What AI Media does	Where to set it
Default providers	Every call uses the AI module default provider and model for its operation; AI Media has no provider setting	Providers
Events and tags	Tags every call ai_media, ai_media_generator:<id>, ai_media_media_type:<bundle> ("Ask AI" adds ai_media_metadata and ai_media_metadata:<target>), and ai_media:<generator> and ai_media_step:<step> when AI Logging or AI Observability is enabled; one request parent ID per Generate press	Providers
Guardrails	Global guardrail sets apply to every call; the Safety tab can attach one of the site's sets. A stop shows the site's message and makes nothing	Settings: Safety
Moderation	None of its own; an AI External Moderation rule with the tag ai_media covers providers without built-in moderation. "Flagged as unsafe" messages make no provider claim	Settings: Safety
AI Prompts	Every prompt it sends (SVG, image, "Ask AI") is an AI Prompt the site can edit	Settings: AI Prompts
AI Context	When AI Context is enabled, an "AI Media" consumer type in Context Control Center: context is pushed into chat requests and pulled into the image prompt	Settings: Site context
AI Logging, AI Observability	Nothing to set: filter on the ai_media, ai_media:<generator> or ai_media_step:<step> tags	Providers
AI Validations	Nothing to set: add an AI image constraint to the media type's image field to check saved images	Settings: Safety
AI Image Alt Text	Its button can fill the alternative text of a generated image	Settings

The Video generator is experimental and off by default: see Providers: Video.