Configuring an embedding connection

This section details how to configure an embedding connection in Kyvos to integrate embedding models (such as OpenAI, Azure OpenAI, AWS Bedrock, or Vertex AI) that convert text into vector embeddings for semantic search and conversational analytics.

An embedding connection in Kyvos refers to the configuration that links the platform to an embedding model—a machine learning model that transforms textual or complex data into dense, numerical vectors called embeddings. This helps improve semantic search and conversational analytics by matching user intent, not just keywords, making it easier to find relevant data using natural language.

You can now specify the type of connection while uploading the custom provider.

After specifying the connection type, you can upload the required JAR and configuration files in a ZIP format corresponding to the selected type.

OpenAI

To configure the GenAI embedding connection, perform the following steps:

Parameter/Field	Description

Parameter/Field	Description
Connection Name	A unique name that identifies your GenAI connections.
Provider	The name of the GenAI provider the system will use to generate output. Select the required provider from the list.
Model	Select the name of the model for generating embeddings.
EndPoint	Specify a unique identifier for the end user, which helps OpenAI monitor and detect abuse. For OpenAI, provide the endpoint in the following format: /chat/completions
Prompt Token Limit	Specify maximum tokens allowed for prompt in single request for current model.
Similarity Upper Threshold	Specify upper threshold to control the selection of records for Retrieval Augmented Generation (RAG).
Similarity Lower Threshold	Specify lower threshold to control Retrieval Augmented Generation (RAG) record selection.
Template Folder Path	Provide the folder path for templates.
RAG Max Records	Specify the maximum number of records (RAG) required in a user prompt.

Azure OpenAI

Parameter/Field	Description

Parameter/Field	Description
Connection Name	A unique name that identifies your GenAI connections.
Provider	The name of the GenAI provider the system will use to generate output. Select the required provider from the list.
Model	Select the name of the model for generating embeddings.
EndPoint	Specify a unique identifier for the end user, which helps OpenAI monitor and detect abuse. For Azure OpenAI, provide the endpoint in the following format: {deployment-id}/embeddings?api-version=2024-10-21 For example, /text-embedding-ada-002/embeddings?api-version=2023-05-15
Prompt Token Limit	Specify maximum tokens allowed for prompt in single request for current model.
Similarity Upper Threshold	Specify upper threshold to control the selection of records for Retrieval Augmented Generation (RAG).
Similarity Lower Threshold	Specify lower threshold to control Retrieval Augmented Generation (RAG) record selection.
Template Folder Path	Provide the folder path for templates.
RAG Max Records	Specify the maximum number of records (RAG) required in a user prompt.

AWS Bedrock

Parameter/Field	Description

Parameter/Field	Description
Connection Name	A unique name that identifies your GenAI connections.
Provider	The name of the GenAI provider the system will use to generate output. Select the required provider from the list.
Embedding Model	Select the name of the model for generating embeddings.
EndPoint	Specify a unique identifier for the end user, which helps OpenAI monitor and detect abuse. For AWS Bedrock, provide the endpoint in the following format: https://bedrock-runtime.{{us-east-1}}.amazonaws.com
Embedding Prompt Token Limit	Specify maximum tokens allowed for prompt in single request for current model.
Similarity Upper Threshold	Specify upper threshold to control the selection of records for Retrieval Augmented Generation (RAG).
Similarity Lower Threshold	Specify lower threshold to control Retrieval Augmented Generation (RAG) record selection.
AWS Region LLM	Provide the region for llm
Template Folder Path	Provide the folder path for templates.
RAG Max Records	Specify the maximum number of records (RAG) required in a user prompt.

VertexAI

Parameter/Field	Description

Parameter/Field	Description
Connection Name	A unique name that identifies your GenAI connections.
Provider	The name of the GenAI provider the system will use to generate output. Select the required provider from the list.
Authentication Type	Select one of the following: ADC: Uses Application Default Credentials available in the environment. Service Account JSON: Uses a service account JSON file. Note: Service Account JSON is only supported for on prem clusters whereas ADC and Service account JSON. Both are supported on GCP Environment.
Location	Specify the Google Cloud region (e.g., us-central1) where Vertex AI resources are hosted and processed.
Project ID	Specifies the unique identifier for your Google Cloud Project where Vertex AI is enabled.
Model	Specify the model name for generating content.
Default Connection For Natural Language Querying	Select this checkbox if you want to make this embedding connection as a default connection.
Prompt Token Limit	Specify maximum tokens allowed for prompt in single request for current model.
Similarity Upper Threshold	Specify upper threshold to control the selection of records for Retrieval Augmented Generation (RAG).
Similarity Lower Threshold	Specify lower threshold to control Retrieval Augmented Generation (RAG) record selection.
RAG Max Records	Specify the maximum number of records (RAG) required in a user prompt.