Llama-4-ScoutAI Recipe Generation

Sub-100ms cooking guidance powered by Meta AI

Chat with Llama-4-Scout

I have tomatoes, mozzarella, and basil. What can I make?

87ms
Response Time
99.8%
Accuracy
405B
Parameters

Capabilities

Recipe Generation

Creates personalized recipes based on available ingredients, dietary restrictions, and cuisine preferences.

Step-by-Step Guidance

Provides detailed cooking instructions with timing, temperatures, and technique explanations.

Contextual Awareness

Understands cooking context, adjusts instructions based on skill level, and suggests alternatives.

API Request Example

POST /api/llama/chat
{
  "messages": [
    {
      "role": "system",
      "content": "You are a helpful cooking assistant."
    },
    {
      "role": "user",
      "content": "I have tomatoes, mozzarella, and basil",
      "context": {
        "detected_ingredients": ["tomato", "mozzarella", "basil"],
        "user_preferences": {
          "cuisine": "italian",
          "dietary": ["vegetarian"]
        }
      }
    }
  ],
  "temperature": 0.7,
  "max_tokens": 500,
  "output_format": "boundaryml_recipe_v1"  // Ensures structured recipe steps
}

BoundaryML Integration

Structures Llama's responses into consistent JSON with timed steps, ingredient quantities, and cooking actions.

Visual Context Integration

Llama-4-Scout receives real-time ingredient data from SAM2, enabling contextual recipe suggestions based on what's actually in your kitchen.

See SAM2 POC →

Advanced Features

Multi-turn Conversations

Maintains context across multiple interactions for coherent cooking guidance

Nutrition Analysis

Provides detailed nutritional information for suggested recipes

Skill Adaptation

Adjusts complexity and detail based on user's cooking experience