The application id of the context entry
OptionalautoNames of metadata fields whose stored values were machine-extracted from the content
(driven by the knowledge base's metadataFields schema, plus the reserved lastModified key
every knowledge base derives from the document container) rather than user-supplied.
On re-upsert/replay, extracted values count as not-supplied (re-extracted fresh) while
user-supplied values are preserved; consumers can also use this to caveat extracted values.
OptionalcontentSHA-256 (hex) of the source bytes (file) or UTF-8 text this context was ingested from. Powers content-level dedup on upsert. Absent on legacy contexts ingested before hashing was introduced.
The date and time the context was created.
The unique identifier of the context entry.
The knowledge base id of the context entry
Metadata associated with the context.
Indicates whether the context is a preview; defaults to false.
OptionalrequestThe options the context was ingested with, replayed when it is reupserted. Absent on contexts that carry none; their reupsert takes the defaults.
The size of the context content in bytes.
The text content of the context.
A title describing the context content.
The type of context (e.g., 'text' or 'file').
The date and time the context was last updated.
Represents an AI knowledge base's context entry with metadata and content.