Skip to content
Litmus

JSON Schema

The schema file should be a valid JSON Schema. It is passed to OpenRouter’s response_format parameter to enforce structured output from the LLM.

Basic Structure

Section titled “Basic Structure”
{
  "type": "object",
  "properties": {
    "field1": { "type": "string" },
    "field2": { "type": "integer" }
  },
  "required": ["field1", "field2"],
  "additionalProperties": false
}

Supported Types

Section titled “Supported Types”

JSON Schema supports several data types:

  • string - Text values
  • integer - Whole numbers
  • number - Decimal numbers
  • boolean - True/false values
  • array - Lists of values
  • object - Nested objects

Examples

Section titled “Examples”

Person Extraction Schema

Section titled “Person Extraction Schema”
{
  "type": "object",
  "properties": {
    "name": { "type": "string" },
    "age": { "type": "integer" },
    "company": { "type": "string" }
  },
  "required": ["name", "age", "company"],
  "additionalProperties": false
}

Sentiment Analysis Schema

Section titled “Sentiment Analysis Schema”
{
  "type": "object",
  "properties": {
    "sentiment": {
      "type": "string",
      "enum": ["positive", "negative", "neutral"]
    },
    "confidence": {
      "type": "number",
      "minimum": 0,
      "maximum": 1
    }
  },
  "required": ["sentiment", "confidence"],
  "additionalProperties": false
}

Entity Extraction with Arrays

Section titled “Entity Extraction with Arrays”
{
  "type": "object",
  "properties": {
    "people": {
      "type": "array",
      "items": {
        "type": "object",
        "properties": {
          "name": { "type": "string" },
          "role": { "type": "string" }
        },
        "required": ["name", "role"]
      }
    },
    "organizations": {
      "type": "array",
      "items": { "type": "string" }
    }
  },
  "required": ["people", "organizations"],
  "additionalProperties": false
}

Best Practices

Section titled “Best Practices”
  1. Use additionalProperties: false - This ensures the LLM only outputs the fields you specify
  2. Mark required fields - Use the required array to specify which fields must be present
  3. Use enums for constrained values - When you need specific values, use enum
  4. Add constraints - Use minimum, maximum, minLength, maxLength for validation

Resources

Section titled “Resources”