ES|QL EMBEDDING function
Elastic Cloud Serverless projects are autoscaled environments, fully managed by Elastic and available on Elastic Cloud. Availability This functionality is in technical preview and is ready for evaluation. Use with caution in production; it is not recommended for mission-critical workloads. Elastic will work to fix any issues, but features in technical preview are not subject to the support SLA of official GA features. Specific Support terms apply. Serverless interfaces and procedures might differ from classic Elastic Stack deployments. The Elastic Stack includes Elastic's core products such as Elasticsearch, Kibana, Logstash, and Beats. Availability This functionality is in technical preview and is ready for evaluation. Use with caution in production; it is not recommended for mission-critical workloads. Elastic will work to fix any issues, but features in technical preview are not subject to the support SLA of official GA features. Specific Support terms apply. Unless stated otherwise on the page, this functionality is available when your Elastic Stack is deployed on Elastic Cloud Hosted, Elastic Cloud Enterprise, Elastic Cloud on Kubernetes, and self-managed environments. This documentation corresponds to the latest patch available for each minor version. If you're not using the latest patch, check the release notes for changes.
Serverless
Preview
Preview
Stack
Preview
9.5+
Preview since 9.5
Generates dense vector embeddings from multimodal input using an inference endpoint.
value- Value to generate embeddings from. Must be a non-null literal string value. Use data type and format options to specify the content type and format (e.g. plain text, base64-encoded images) of the input.
inference_id- Identifier of an existing inference endpoint that will generate the embeddings. The inference endpoint must have the
embeddingtask type and should use the same model that was used to embed your indexed data. options-
(Optional) Options for the input value.
Generates dense vector embeddings from multimodal input using a specified inference endpoint with the embedding task type. Use this function to generate query vectors for KNN searches from multimodal inputs against your vectorized data or other dense vector based operations.
| value | inference_id | options | result |
|---|---|---|---|
| keyword | keyword | named parameters | dense_vector |
| keyword | keyword | dense_vector |
type- (keyword) Content type of the input (e.g. "text", "image").
timeout-
(keyword) Timeout for the inference request (e.g. "30s", "1m").
Generate embeddings using the 'test_dense_inference' inference endpoint, using the default type and format (text):
FROM dense_vector_text METADATA _score
| EVAL query_embedding = EMBEDDING("be excellent to each other", "test_embedding_inference")
| WHERE KNN(text_embedding_field, query_embedding)
| SORT _score DESC
| LIMIT 10
Generate embeddings using an inference endpoint, specifying the data type:
ROW b64_input="This is an image"
| EVAL embedding = EMBEDDING("data:image/jpeg;base64,VGhpcyBpcyBhbiBpbWFnZQ==", "test_embedding_inference", {"type": "image"})