Skip to main content
Top
Review meshIQ Products (G2.com)

Add and Manage AI Agent Models

The AI Agent Settings window is where administrators register the large language models (LLMs) that the meshIQ Agent uses. Each entry defines a provider, a model, its connection details and its generation settings. Administrators can also designate a model as the default for the meshIQ Agent.

Access AI Agent Settings

  1. Click the Settings SttgsIcon.png icon in the top-right toolbar.

  2. Navigate to Application Settings.

  3. Select AI Agent Settings.

Adding a Model

Prerequisites

  • Provider and Model Details: The provider name and the exact model identifier expected by the provider.
  • API Key: A valid API key for the selected provider, if authentication is required.
  • Base URL: For self-hosted or OpenAI-compatible endpoints, the base URL must be reachable from the meshIQ Manage server.

Procedure

  1. In the AI Agent Settings window, click the Add Add buttin.png icon in the toolbar. The Add Model panel opens on the right.

  2. In Name, replace New Model with a descriptive name for the model entry.

    Turn on Make it Default for MeshIQ Agent if you want the meshIQ Agent to use this model by default.

  3. Select a Provider from the drop-down list.
  4. In Model, enter the exact model identifier expected by the provider. For example claude-opus-4-6, gpt-5 or llama3.1:8b.
  5. Enter the API Key, if the provider requires authentication.
  6. Enter the Base URL if you selected OpenAI-compatible (self-hosted / OSS) or are configuring a local endpoint.
  7. Configure Temperature, Max tokens, and Tool-calling mode, as needed, or retain their default values.

  8. Click Save. The model appears in the AI Agent Settings list.

Field descriptions

FieldRequiredDescription
NameYesDisplay name for this model entry in the AI Agent Settings list. Use a name that identifies the provider, model and purpose.
Make it Default for MeshIQ AgentNoWhen enabled, the meshIQ Agent uses this model by default. Only one model can be set as the default, if supported by the application configuration.
ProviderYesThe LLM provider or API type used to access the model. Select OpenAI-compatible (self-hosted / OSS) for self-hosted, open-source, or other endpoints that implement the OpenAI API.
ModelYesThe exact model identifier expected by the provider.
API KeyDepends on providerThe secret key used to authenticate with the provider. Leave this field blank for local endpoints that do not require authentication.
Base URLDepends on providerThe endpoint address, starting with https:// (or http:// for a local server). Required for OpenAI-compatible and locally hosted models.
TemperatureNoControls the randomness of responses. Lower values produce more consistent responses, while higher values produce more varied output. Use a lower value for operational tasks when consistency is important.
Max tokensNoSpecifies the maximum number of tokens the model can generate in a single response. Increase this value if responses are truncated, without exceeding the model's output limit.
Tool-calling modeNo

How the agent invokes meshIQ tools through the model. 

Available Values:

  • Native (function calling): To use the model's built-in function-calling API. 
  • Prompted (ReAct): To invoke tools through prompt-based reasoning.
  • None (chat only): To disable tool calling and enable chat-only interactions.

Managing existing models

  • Edit a model: select the model from the list and click the Edit Selected icon, Update the required fields and click Save.
  • Delete a model: Select the model from the list and click the Delete Selected icon. If you delete the default model, set another model as the default.
  • Find a model: Enter part of the model's name in the Search box to find the matching entry.