AI Models

Use AI Model settings to control how generation and evaluation prompts are executed.

Provider guides

• OpenAI
• Azure OpenAI
• Claude (Anthropic)
• Gemini (Google)
• Ollama
• Disabled/Mock providers

What this controls

• Provider/backend selection.
• Generation and evaluation model behaviour.
• Prompt execution defaults used in Workflow Outputs and Evaluation tabs.

How to use it

1. Open Settings > AI Model.

2. Configure the backend/provider required in your environment.

3. Set generation/evaluation defaults for your workflow.

4. Run a small test prompt in Workflow before large runs.

Practical tips

• Keep one stable baseline model for repeatable comparisons.
• If output quality shifts, verify model/backend settings before changing policies.