Sanity Library Reference Docs
    Preparing search index...

    Interface AgentActionTargetBeta

    interface AgentActionTarget {
        exclude?: AgentActionPathSegment[];
        maxPathDepth?: number;
        path?: AgentActionPathSegment | AgentActionPath;
        types?: AgentActionTypeConfig;
    }

    Hierarchy (View Summary)

    Index

    Properties

    By default, all children up to target.maxPathDepth are included. Fields or array items not on the exclude list, are implicitly included.

    maxPathDepth?: number

    maxPathDepth controls how deep into the schema from the target root the instruction will affect.

    Depth is based on path segments:

    • title has depth 1
    • array[_key="no"].title has depth 3

    Be careful not to set this too high in studios with recursive document schemas, as it could have negative impact on performance; both for runtime and quality of responses.

    Default: 4

    Root target path.

    Use this to have the instruction only affect a part of the document.

    To further control the behavior of individual paths under the root, use include, exclude, types.include and types.exclude.

    Example:

    path: ['body', {_key: 'someKey'}, 'nestedObject']

    Here, the instruction will only write to fields under the nestedObject.

    Default: [] = the document itself

    • #AgentActionPathSegment
    • #AgentActionPath

    Types can be used to exclude array item types or all fields directly under the target path of a certain type. If you do exclude: ['string'] all string fields under the target will be excluded, for instance.

    types.include and types.exclude are mutually exclusive.