Vector Index & Attribute Settings
Configure the vector dimensionality, search thresholds, HNSW tuning, and which catalog attributes become embeddings.
Vector Index & Attribute Fields
| Field | Description |
|---|---|
| AI Model Dimension | Vector dimensionality matching the embedding model (e.g. 1536 for OpenAI, 768 or 1024 for Gemini, 384 for Intfloat). |
| Minimum Searching Score | Minimum similarity score threshold for catalog search results (value > 0). |
| HNSW Index M Value | (Default Storage only) Number of bidirectional links per node in HNSW graph (Range: 4–64). Higher values improve search accuracy at the cost of indexing speed. |
| HNSW Index EF Construction Quality | (Default Storage only) Dynamic candidate list size during indexing (Range: 50–1000). Higher values improve search recall. |
| Select Attributes for Vector Storage | Multi-select catalog attributes included when creating embeddings (e.g., Name, SKU, Description, Short Description, Price, Color, Material, Category). |
| Regenerate Embeddings On Product Save | When enabled (Yes), saving a product in Magento automatically triggers a background queue task to update its embedding. |
| Admin Chat Profile Image | Upload custom avatar image displayed for the chatbot assistant in the chat header. |
| ChatBot Name | Custom name displayed for the chatbot assistant (e.g., "Store Assistant"). |