Options
All
  • Public
  • Public/Protected
  • All
Menu

EmbeddingsPage

Returned by JoplinAi.getEmbeddings.

The vectors are produced by a specific model: comparing vectors across different modelId values (or persisting vectors across model changes) is meaningless. Plugins that cache vectors must key the cache by modelId and invalidate when it changes between calls.

Index

Properties

chunks

chunks: EmbeddingChunk[]

dimension

dimension: number

Length of each vector array.

modelId

modelId: string

Identifier of the model that produced these vectors, e.g. 'multilingual-e5-small'.

Optional nextCursor

nextCursor: string

Pass this back as cursor on the next call to get the following page. undefined means there are no more results in the current snapshot.

If the index is wiped between pages (e.g. the user changes the embedding model), pagination ends silently — the next call with this cursor returns an empty chunks array. Plugins should watch for a modelId change between pages and discard partial results if it differs.