For the complete documentation index, see llms.txt. This page is also available as Markdown.

Metadata Enrichment Agent

Metadata Enrichment Agent is a Visual AI feature that allows you to automatically generate metadata values using an AI prompt.

Metadata Enrichment Agent is compatible with 4 types of metadata :

  • Boolean

  • Number (integer)

  • List (single value)

  • List (multiple values)

For a metadata field to be automatically filled by AI Agent you need :

  • The metadata field to have an advanced field named ai_prompt, set your prompt as value and choose a compatible metadata type.

  • Set on a Agent: AI Enrichment post-process.

Create a Metadata Enrichment Agent compatible field

Go in the DAM metadata configuration page to add a new metadata field :

  • Choose a compatible metadata type (see above).

  • Add an advanced field named ai_prompt and type your prompt as value. In the following example we are creating a boolean metadata (True/False) that will indicate if any children is visible on the image using the AI prompt : Is there any child visible on the image ?.

Create an AI enrichment post-process

Go to the DAM post-process configuration page and add a new post-process Agent: AI enrichment.

Notes

  • After creating the post-process all compatible metadata that have an AI prompt will automatically be filled after upload.

  • This is a DAM feature, it works only when we can retrieve metadata configuration from token’s settings.

  • For now there is no limitation on the amount of AI compatible metadata.

  • Demo video here.

Last updated