Articles in this section

Configuring AI in QuerySurge (Versions: 14.2+)

Overview

QuerySurge is evolving with powerful AI integrations designed to transform how teams validate and interact with data. To leverage this innovation, you'll need to configure your global AI engine.

  1. QuerySurge AI Core — an on-premises engine that brings advanced AI capabilities directly into your existing QuerySurge environment.
  2. QuerySurge AI Cloud — a cloud-based service that uses OpenAI to unlock greater processing power and flexibility without the local infrastructure requirements.

Together, these options give your organization the freedom to choose the deployment model that best fits their security, performance, and scalability needs.

Configuration

In the Global Administration portal, navigate to Artificial Intelligence → Configuration and check the Enable Artificial Intelligence checkbox.

config-core-default.png

Next, choose your desired Platform. Select QuerySurge AI Core if you have a locally hosted deployment; otherwise, select QuerySurge AI Cloud. There is a single Provider for each platform — Core for QuerySurge AI Core and OpenAI for QuerySurge AI Cloud.

Please note, only platforms for which you're licensed will be available in the dropdown menu.

QuerySurge AI Core

For a local deployment of QuerySurge AI, you'll need the following information.

  • Scheme — either HTTP or HTTPS (this depends on whether your AI server server supports SSL)
  • Hostname — the fully qualified domain name of your AI server
  • Port — the port number on which your AI server is listening
  • Authentication Timeout — how often your application server will re-authenticate with your AI server
config-core-edited.png

Advanced Model Settings

There are three (3) advanced model settings that allow you to tune the model itself.

  • Threads — the number of threads to use during SQL generation (generally speaking, we recommend setting this value equal to the number of physical CPU cores available on your AI server)
  • Context Size — the size (in tokens) of the context window for your model (for models supporting a larger context, it might make sense to increase this value accordingly)
  • Max. Tokens — the maximum number of tokens allowed in a generated response (increasing this value may improve performance, but it'll be at the expense of added memory usage)

Please note, your QuerySurge AI Core application settings must match your QuerySurge AI Core server settings.

QuerySurge AI Cloud

Cloud deployments of QuerySurge AI utilize OpenAI for SQL generation. That being said, you must have your own OpenAI credentials or have access to QuerySurge-bundled credentials as part of your license.

The setup process is similar, and requires a lot of the same information as QuerySurge AI Core.

  • Scheme — either HTTP or HTTPS (by default, this is HTTPS)
  • Hostname — the address of the OpenAI server (by default, this is api.open.com)
  • Port — the port number on which your AI server is listening (by default, this is 443)
  • Path — the OpenAI REST API version (by default, this is /v1)
  • Model — the model you wish to use for SQL generation
config-cloud-default.png

Supported Models

We support the following models, which are accessible in the Model dropdown menu.

  • gpt-5
  • gpt-5-chat
  • gpt-5-mini
  • gpt-5-nano
  • gpt-4.1
  • gpt-4.1-mini
  • gpt-4.1-nano
  • o4-mini
  • o4-nano

Please note, if you're using your own credentials you must verify that you have access (in your OpenAI account) to the model you choose. If you're using QuerySurge-bundled credentials, you will have access to any model available in the Model dropdown menu.

Credentials

To use QuerySurge AI Cloud, you must have valid OpenAI credentials. If you have access to this through your organization, enter your API Key, and (optionally) your Organization ID and Project ID.

config-cloud-edited.png

Otherwise, and if licensed to do so, toggle the Use QuerySurge credentials? button to use the OpenAI credentials bundled with your license.

config-cloud-bundled.png

Please note, the Scheme, Hostname, Port, Path, API Key, Organization ID, and Project ID are readonly when using QuerySurge-bundled credentials.

Advanced Model Settings

There is only one (1) setting to tune the model.

  • Max. Tokens — the maximum number of tokens allowed in a generated response

Validating your Settings

You can (optionally) validate your connection details with either QuerySurge AI Core or QuerySurge AI Cloud. This sends a small, simple prompt to the model to validate connectivity with the server and readiness of the model.

configuration-06.png

When you're ready, click the Save button to save your settings. You can now start leveraging the power of AI to build data validation tests in QuerySurge.

 

 

Related articles