There are two approaches for retrieving resources: filter and search.

For simple, classic retrieval of content, filtering is probably adequate and is driven by metadata. For fuzzier, content-driven, or semantic retrieval, use the search endpoint.

Filter/get resources

/v2/resources/filter endpoint enables you to list all resources with a specific ordering-criteria (e.g. created at, modified at).

Tip: by providing the parentId property, the endpoint will list all direct child resources of the given parent resource id.

Searching resources

/v2/search allows you to search resources with content-driven tooling and supplementary filters. The default ordering of resources is based on highest-to-lowest relevancy.

Fabric’s search is hybrid, combining keyword and semantic search.

• Expected input: text

• Retrieval is based on a fusion of contained text (e.g. text content of a note) and semantic retrieval based on a generated summary embedding. For images, retrieval is also based on the embedding of the image itself.

• The search can be coerced to find a strict string match by wrapping any part of the text query in double quotes.

Search filters:

Search supports additional filtering. Filters can include color, tags, the kind of resource and more. Filters work by excluding non-matching results, and in the case of color, further acts as a relevancy factor, prioritising closer color matches.