Structured outputs

When you need machine-readable output, ask for JSON with response_format. Two modes: plain JSON mode, where the model must return valid JSON; and JSON schema, where the output is constrained to a shape you define. Both use the OpenAI response_format field.

JSON mode

Set response_format to {"type": "json_object"} and the model returns syntactically valid JSON. Describe the fields you want in your prompt; JSON mode guarantees the output parses, not that it matches a particular shape.

curl https://api.merius.ai/v1/chat/completions \
  -H "Authorization: Bearer $MERIUS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "z-ai/glm-5.1",
    "response_format": {"type": "json_object"},
    "messages": [
      {"role": "user", "content": "Return name and city for Ada Lovelace as JSON."}
    ]
  }'

JSON schema

For a guaranteed shape, pass a json_schema. The model's output is constrained to your schema, so every field you mark required is present and correctly typed — no post-validation needed for structure:

"response_format": {
  "type": "json_schema",
  "json_schema": {
    "name": "person",
    "schema": {
      "type": "object",
      "properties": {
        "name": {"type": "string"},
        "city": {"type": "string"}
      },
      "required": ["name", "city"]
    }
  }
}