Magic AI Configuration

The Magic AI section in the admin sidebar provides four sub-pages for managing AI providers, behavior, prompts, and system-level personality settings. Navigate to Magic AI in the sidebar to access them.

Platforms

Navigate to Magic AI >> Platforms to manage your AI provider connections.

The Platforms page displays a datagrid with the following columns:

Column	Description
Label	The name you assigned to the platform configuration
Provider	The AI provider (OpenAI, Gemini, Ollama, Groq, etc.)
Models	The models enabled for this platform
Default	Whether this platform is the default (Yes/No)
Status	Enabled or Disabled
Created At	Date the platform was added
Actions	Star (set as default), Edit (pencil icon), Delete (trash icon)

Adding a Platform

Click the Add Platform button in the top-right corner. A modal opens with the following fields:

1. Provider — Select a provider from the dropdown (e.g., OpenAI, Gemini, Ollama, Groq).
2. Label — Enter a descriptive name for this platform configuration.
3. API Key — Paste the API key from your provider account.
4. Models — A multiselect field listing available models for the selected provider. Choose one or more models.
5. Status — Toggle to enable or disable the platform.

TIP

API credentials are stored with encrypted credential storage for security. Your API keys are never stored in plain text.

Platform Actions

• Star icon — Sets the platform as the default. Only one platform can be the default at a time.
• Pencil icon — Opens the edit modal to update the platform label, API key, models, or status.
• Trash icon — Deletes the platform configuration. This action cannot be undone.

Settings

Navigate to Magic AI >> Settings in the sidebar. This opens the configuration page at /admin/configuration/general/magic_ai with four sections.

1. Agentic PIM

Configure the AI Agent Chat, autonomous enrichment workflows, and quality monitoring.

Field	Description
Enable AI Agent Chat	Toggle to enable or disable the conversational AI agent interface
Max Agent Steps Per Turn	Dropdown to set the maximum number of tool calls the agent can make per turn (default: 5)
Daily Token Budget	Number field to set the maximum tokens the AI can consume per day (e.g., 500000)
Auto-Enrichment on Product Create	Toggle to automatically enrich product data when a new product is created
Catalog Quality Monitor	Toggle to enable ongoing AI-driven catalog quality monitoring
Confidence Threshold	Dropdown to set how confident the AI must be before applying changes (default: 0.7 Balanced)
Change Approval Mode	Dropdown to control how AI-proposed changes are applied (default: "Confirm & apply" — the AI proposes values, asks for confirmation, then executes)

2. Text Generation

Configure the default AI platform and model for generating product descriptions, category content, and other text-based AI features.

Field	Description
Enabled	Toggle to enable or disable text generation
Default Platform	Dropdown to select the platform. Choose "Use Default Platform" to use the starred platform, or select a specific one
Default Model	Dropdown to select the model from the chosen platform

3. Image Generation

Configure the default AI platform and model for generating product images. Only platforms that support image generation (OpenAI, Gemini, xAI) are listed.

Field	Description
Enabled	Toggle to enable or disable image generation
Default Platform	Dropdown to select an image-capable platform
Default Model	Dropdown to select the image generation model

4. Translation

Configure the AI platform and model for translating product content across locales. You can use a different platform for translations, allowing you to choose a faster or more cost-effective provider.

Field	Description
Enabled	Toggle to enable or disable AI-powered translation
Default Platform	Dropdown to select the platform used for translation
Translation Model	Dropdown to select the model used for translation
Replace Existing Value	Toggle to control whether existing translations are overwritten when re-translating
Source Channel	Dropdown to select the channel whose content is used as the translation source
Target Channel	Dropdown to select the channel that receives translated content
Source Locale	Dropdown to select the locale to translate from
Target Locales	Multi-select dropdown to choose which locales to translate into

TIP

You can assign a different (potentially cheaper or faster) AI provider specifically for translations, keeping your premium provider for content generation.

Click Save Configuration at the bottom of the page to apply all changes.

Prompts

Navigate to Magic AI >> Prompts to manage custom prompts for AI content generation.

The Prompts page displays a datagrid with the following columns:

Column	Description
Title	The name of the prompt
Prompt	The prompt text with placeholders
Entity Type	The entity the prompt applies to (product or category)
Purpose	Whether the prompt is for Text Generation or Image Generation
Created At	Date the prompt was created
Updated At	Date the prompt was last modified
Actions	Edit (pencil icon), Delete (trash icon)

Creating a Prompt

Click the Create Prompt button to add a new prompt. Define the title, prompt text, entity type, and purpose.

Prompts support dynamic placeholders that are replaced with actual product or category data at generation time. Common placeholders include @name, @color, and other attribute codes.

Example Prompts

Here are examples of prompts that ship with UnoPim:

Title	Prompt	Entity Type	Purpose
AI Product Description	Write a detailed product description for @name highlighting its features, benefits and @color variant.	product	Text Generation
AI Product Image	Generate a professional product photo of @name on a clean white background with studio lighting.	product	Image Generation
AI Category Description	Write a compelling category description for @name that helps customers browse products.	category	Text Generation

TIP

Use attribute codes as placeholders (prefixed with @) in your prompts. The AI will replace them with actual values from the product or category being processed.

System Prompts

Navigate to Magic AI >> System Prompts to configure the AI's personality and behavior for content generation.

The System Prompts page displays a datagrid with the following columns:

Column	Description
Title	The name of the system prompt
Tone	The conversational tone (e.g., Confident, Vivid, Brief)
Max Tokens	The maximum number of tokens for AI responses
Temperature	The creativity level (lower = more focused, higher = more creative)
Status	Enabled or Disabled
Created At	Date the system prompt was created
Updated At	Date the system prompt was last modified
Actions	Edit (pencil icon), Delete (trash icon)

Preset System Prompts

UnoPim ships with 10 preset system prompts. Only one system prompt can be enabled at a time.

Title	Tone	Temperature	Notes
Authoritative Guide	Confident, assertive, instructional	0.65
Descriptive Storyteller	Vivid, rich, engaging	0.9
Concise Responder	Brief, to-the-point	0.5
Technical Expert	Precise, analytical	0.6
Casual Conversationalist	Informal, relaxed	0.75
Motivational Coach	Energetic, encouraging	0.85
Empathetic Listener	Warm, understanding	0.6
Witty Commentator	Clever, humorous	0.9
Professional Advisor	Formal, respectful	0.65
Friendly Assistant	Friendly, helpful, casual	0.7	Enabled by default

Creating a System Prompt

Click the Create System Prompt button to define a new AI personality. Configure the title, tone description, max tokens, temperature, and status.

TIP

Only one system prompt can be active at a time. Enabling a new system prompt automatically disables the previously active one. Choose a system prompt that matches the tone you want across all AI-generated content.