llm/docs/schemas.md

(schemas)=

Schemas

Large Language Models are very good at producing structured output as JSON or other formats. LLM's schemas feature allows you to define the exact structure of JSON data you want to receive from a model.

This feature is supported by models from OpenAI, Anthropic, Google Gemini and others {ref}via plugins <advanced-model-plugins-schemas>.

Understanding JSON schemas

A JSON schema is a specification that describes the expected structure of a JSON object. It defines:

• The data types of fields (string, number, array, object, etc.)
• Required vs. optional fields
• Nested data structures
• Constraints on values (minimum/maximum, patterns, etc.)
• Descriptions of those fields - these can be used to guide the language model

Different models may support different subsets of the overall JSON schema language. You should experiment to figure out what works for the model you are using.

Using schemas with LLM

LLM provides several ways to use schemas:

1. Directly via the command line with the --schema option
2. Through stored schemas in the database
3. Via templates that include schemas
4. Through the {ref}Python API <python-api-schemas>

Basic usage with the command line

To get structured data from a language model you can provide a JSON schema directly using the --schema option:

curl https://www.nytimes.com/ | uvx strip-tags | llm --schema '{
  "type": "object",
  "properties": {
    "stories": {
      "type": "array",
      "items": {
        "type": "object",
        "properties": {
          "headline": {
            "type": "string"
          },
          "short_summary": {
            "type": "string"
          },
          "key_people": {
            "type": "array",
            "items": {
              "type": "string"
            }
          }
        },
        "required": ["headline", "summary", "key_people"]
      }
    }
  },
  "required": ["stories"]
}' | jq

This example uses uvx to run strip-tags against the front page of the New York Times, runs GPT-4o mini with a schema to extract story headlines and summaries, then pipes the result through jq to format it.

This will instruct the model to return an array of JSON objects with the specified structure, each containing a headline, summary, and array of key people mentioned.

Alternative schema syntax

JSON schema's can be time-consuming to construct by hand. LLM also supports a concise alternative syntax for specifying a schema.

A simple schema for an object with two string properties called name and bio looks like this:

name, bio

You can include type information by adding a type indicator after the property name, separated by a space.

name, bio, age int

Supported types are int for integers, float for floating point numbers, str for strings (the default) and bool for true/false booleans.

To include a description of the field to act as a hint to the model, add one after a colon:

name: the person's name, age int: their age, bio: a short bio

If your schema is getting long you can switch from comma-separated to newline-separated, which also allows you to use commas in those descriptions:

name: the person's name
age int: their age
bio: a short bio, no more than three sentences

This format is supported by the --schema option. The format will be detected any time you provide a string with at least one space that doesn't start with a { (indicating JSON):

llm --schema 'name,description,fave_toy' 'invent a dog'

To return multiple items matching your schema, use the --schema-multi option. This is equivalent to using --schema with a JSON schema that specifies an items key containing multiple objects.

llm --schema-multi 'name,description,fave_toy' 'invent 3 dogs'

The Python utility function llm.utils.build_json_schema(schema) can be used to convert this syntax into the equivalent JSON schema dictionary.