Contents

AI settings - Plugixa Recipe

The AI tab configures the LLM provider used by the AI Generator - API key, model, and generation tuning.

Go to Recipes → Settings → AI.

The AI settings tab: an Enable AI features switch, an AI Provider choice of OpenAI or Google Gemini, a masked API key field with Load Models and Test Connection buttons, a model selector, and generation settings for temperature, max tokens and default language

Nutrition is not estimated automatically. There is no setting that fills a nutrition label on its own. The AI Generator has a separate Estimate nutrition action that reads the ingredients and the servings count and returns a per-serving label for you to review before saving.

Options

Key Default Notes
ai_enabled false Master switch. When off, the AI Generator menu item is hidden.
ai_provider openai openai or gemini (Google AI Studio).
ai_openai_api_key (empty) OpenAI key. Stored in plugixa_recipe_settings and masked in the UI once saved.
ai_google_api_key (empty) Google AI Studio key, used when ai_provider is gemini.
ai_model_openai gpt-4o-mini OpenAI model, e.g. gpt-4o or gpt-4o-mini.
ai_model_google gemini-1.5-flash Gemini model, e.g. gemini-1.5-flash or gemini-1.5-pro.
ai_max_tokens 4000 Output cap. The generator auto-scales this for “expanded” prompts; values below 1500 may truncate nutrition and instruction JSON.
ai_temperature 0.7 0.0–1.0. Lower is more factual, higher is more creative.
ai_default_language en Language code for LLM output. Overridden per-generation in the AI Generator UI.

Getting an API key

  • OpenAI - sign up at platform.openai.com, create a secret key, and paste it. The gpt-4o-mini model costs roughly $0.15/1M input tokens - a typical recipe generation is ~$0.002.
  • Google Gemini - get a free key at aistudio.google.com. gemini-1.5-flash is free for generous daily quotas.

Click Test Connection to verify the key without burning a full generation. Load Models queries the provider for the list of models your key can access.

Usage & quota

Each generation returns a completion cost in the response. Track total usage in your provider’s billing dashboard - Plugixa does not meter or cap spend itself.

Quick Links

Plugin page Changelog