Default Models
PrxmptStudix uses AI to power many of its internal features—from refining your prompt templates to generating entire business profiles. The Default Models page allows you to control which specific AI model is used for each task.
1. Feature Assignments
You can choose to use a single "Universal" model for everything or assign specialized models to specific feature sets.
Universal Default
Function: Enables a single model to handle all internal AI tasks.
Best For: Simplicity and cost management. If you have a preferred "workhorse" model (like GPT-4o or Claude 3.5 Sonnet), this is the easiest configuration.
Granular Assignments: If the "Use Same for All Features" switch is disabled, you can pick different models for:
Prompt Library: Refining, summarizing, or tagging prompt templates.
Persona Generator: Creating detailed demographics and psychographics for AI Personas.
Business Generator: Generating tags, slogans, and branding profiles for Business records.
2. Configuration Logic
Provider First: To select a model, you must first choose an active Provider (e.g., OpenAI).
Automatic Fallback: If a previously assigned model becomes unavailable (e.g., its provider is disabled), the system will prompt you to reassign it.
Requirement: You must have at least one AI Provider configured with a valid API key to use any of these features.
3. Best Practices
Cost vs. Quality: Use high-reasoning models (Like Claude Opus or GPT-4) for the Persona Generator where nuance is critical, while using faster, cheaper models (Like GPT-3.5 or Claude Haiku) for library tagging.
Local Models: If you use Ollama, you can assign local models as your defaults to handle internal generations entirely offline, ensuring maximum privacy for your proprietary brand data.