Interface: MultimodalIndexerConfig

Defined in: packages/agentos/src/cognition/rag/multimodal/types.ts:370

Configuration for the MultimodalIndexer.

Example

const config: MultimodalIndexerConfig = {
  defaultCollection: 'knowledge-base',
  imageDescriptionPrompt: 'Describe this image for use in a search index.',
};

Properties

defaultCollection?

optional defaultCollection: string

Defined in: packages/agentos/src/cognition/rag/multimodal/types.ts:375

Default vector store collection name for indexed content.

Default

'multimodal'

---

imageDescriptionPrompt?

optional imageDescriptionPrompt: string

Defined in: packages/agentos/src/cognition/rag/multimodal/types.ts:381

Custom prompt template for the vision LLM when describing images. If omitted, a sensible default prompt is used.