AI Search - Custom metadata filtering for AI Search
Key Points
- Custom metadata filtering now available for AI Search
- Support for up to 5 custom fields per instance
- Advanced query filtering with comparison operators
Summary
Cloudflare AI Search now supports custom metadata filtering, enabling developers to define custom metadata fields and filter search results based on specific attributes like category, version, or any user-defined field.
Key Points
- Custom metadata schema: Define up to 5 custom metadata fields per AI Search instance with support for
text,number, andbooleandata types - Flexible metadata attachment: Add metadata via S3-compatible headers for R2 buckets or HTML meta tags for websites
- Advanced filtering: Use custom metadata fields alongside built-in attributes like folder and timestamp in search queries
- Query operators: Support for comparison operators like
$gtefor numeric fields in filter expressions - API integration: Full REST API support for creating instances with custom metadata schemas and performing filtered searches