Advanced Fabric Settings

Take your fabrics to the next level with advanced configuration options.

Input Types Deep Dive

Fabrics support 13+ input types, each with specific validation options:

Text Inputs

{
  "type": "text",
  "label": "Blog Topic",
  "placeholder": "Enter your topic",
  "minLength": 10,
  "maxLength": 500,
  "required": true
}

Select Inputs

{
  "type": "select",
  "label": "Tone",
  "options": ["professional", "casual", "humorous"],
  "default": "professional"
}

Slider Inputs

{
  "type": "slider",
  "label": "Creativity",
  "min": 0,
  "max": 100,
  "step": 10,
  "default": 50
}

Image Inputs

{
  "type": "image",
  "label": "Reference Image",
  "maxSize": "10MB",
  "formats": ["png", "jpg", "webp"]
}

Model Configuration

Temperature & Sampling

• Temperature (0-2): Lower = more focused, higher = more creative
• Top P: Controls diversity of outputs
• Max Tokens: Limit output length

Model Selection

Choose models based on task:

• Claude Sonnet: Best for nuanced writing
• GPT-4o: Strong reasoning and coding
• Gemini Flash: Fast, cost-effective
• DeepSeek: Budget-friendly option

Output Processing

Post-Processing Options

• Trim whitespace: Clean up output
• Parse JSON: Structured data extraction
• Format markdown: Render formatting

Output Validation

{
  "outputType": "json",
  "schema": {
    "type": "object",
    "properties": {
      "title": { "type": "string" },
      "summary": { "type": "string" }
    },
    "required": ["title", "summary"]
  }
}

Conditional Logic

Show/hide fields based on other inputs:

{
  "conditionalFields": [
    {
      "field": "customInstructions",
      "showWhen": {
        "field": "useCustomPrompt",
        "value": true
      }
    }
  ]
}

Best Practices

1. Validate early: Use input validation to catch errors
2. Default wisely: Set sensible defaults for optional fields
3. Document fields: Add help text for complex inputs
4. Test thoroughly: Try edge cases before publishing
5. Version control: Use versioning for major changes

Monitoring & Analytics

Track fabric performance:

• Execution time per run
• Credit usage patterns
• Error rates and types
• User satisfaction (ratings)

Learn more: Fabric Chains | API Documentation