> ## Documentation Index
> Fetch the complete documentation index at: https://docs.arouter.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Structured Outputs

> Force models to output valid JSON matching an exact schema. Use json_schema mode for guaranteed schema compliance.

Structured Outputs force the model to return a JSON object that matches a schema you define. This eliminates the need to parse, validate, or retry unstructured model output.

ARouter supports two structured output modes:

| Mode          | Description                                  |
| ------------- | -------------------------------------------- |
| `json_object` | Guarantees valid JSON; no schema enforcement |
| `json_schema` | Guarantees JSON matching your exact schema   |

## Using Structured Outputs

Pass `response_format` in your request body:

### JSON Object Mode

```json theme={null}
{
  "model": "openai/gpt-5.4",
  "messages": [
    {
      "role": "user",
      "content": "Return information about the planet Mars as JSON"
    }
  ],
  "response_format": { "type": "json_object" }
}
```

The model returns valid JSON, but the schema is not enforced:

```json theme={null}
{
  "choices": [
    {
      "message": {
        "role": "assistant",
        "content": "{\"name\":\"Mars\",\"diameter_km\":6779,\"moons\":2,\"habitable\":false}"
      }
    }
  ]
}
```

### JSON Schema Mode

Use `json_schema` to enforce a strict schema:

```json theme={null}
{
  "model": "openai/gpt-5.4",
  "messages": [
    {
      "role": "system",
      "content": "You are a data extraction assistant."
    },
    {
      "role": "user",
      "content": "Extract the planet information: Mars is a terrestrial planet with 2 moons and a diameter of 6779 km."
    }
  ],
  "response_format": {
    "type": "json_schema",
    "json_schema": {
      "name": "planet_info",
      "strict": true,
      "schema": {
        "type": "object",
        "properties": {
          "name": { "type": "string" },
          "diameter_km": { "type": "number" },
          "moons": { "type": "integer" },
          "habitable": { "type": "boolean" }
        },
        "required": ["name", "diameter_km", "moons", "habitable"],
        "additionalProperties": false
      }
    }
  }
}
```

Response:

```json theme={null}
{
  "choices": [
    {
      "message": {
        "role": "assistant",
        "content": "{\"name\":\"Mars\",\"diameter_km\":6779,\"moons\":2,\"habitable\":false}"
      },
      "finish_reason": "stop"
    }
  ]
}
```

## Full Example

<Tabs>
  <Tab title="Python (OpenAI)">
    ```python theme={null}
    import json
    from openai import OpenAI
    from pydantic import BaseModel

    client = OpenAI(
        base_url="https://api.arouter.ai/v1",
        api_key="lr_live_xxxx",
    )

    # Define your schema using Pydantic
    class PlanetInfo(BaseModel):
        name: str
        diameter_km: float
        moons: int
        habitable: bool

    # Using OpenAI's parse() method (recommended)
    response = client.beta.chat.completions.parse(
        model="openai/gpt-5.4",
        messages=[
            {"role": "system", "content": "You are a data extraction assistant."},
            {"role": "user", "content": "Tell me about Mars."},
        ],
        response_format=PlanetInfo,
    )

    planet = response.choices[0].message.parsed
    print(planet.name)        # "Mars"
    print(planet.moons)       # 2
    print(planet.diameter_km) # 6779.0

    # Or manually using response_format dict
    response = client.chat.completions.create(
        model="openai/gpt-5.4",
        messages=[
            {"role": "user", "content": "Tell me about Mars as JSON."},
        ],
        response_format={
            "type": "json_schema",
            "json_schema": {
                "name": "planet_info",
                "strict": True,
                "schema": PlanetInfo.model_json_schema(),
            },
        },
    )
    data = json.loads(response.choices[0].message.content)
    planet = PlanetInfo(**data)
    ```
  </Tab>

  <Tab title="Node.js (OpenAI)">
    ```typescript theme={null}
    import OpenAI from "openai";
    import { zodResponseFormat } from "openai/helpers/zod";
    import { z } from "zod";

    const client = new OpenAI({
      baseURL: "https://api.arouter.ai/v1",
      apiKey: "lr_live_xxxx",
    });

    // Define schema with Zod
    const PlanetInfo = z.object({
      name: z.string(),
      diameter_km: z.number(),
      moons: z.number().int(),
      habitable: z.boolean(),
    });

    // Using parse() with Zod (recommended)
    const response = await client.beta.chat.completions.parse({
      model: "openai/gpt-5.4",
      messages: [
        { role: "system", content: "You are a data extraction assistant." },
        { role: "user", content: "Tell me about Mars." },
      ],
      response_format: zodResponseFormat(PlanetInfo, "planet_info"),
    });

    const planet = response.choices[0].message.parsed;
    console.log(planet?.name);        // "Mars"
    console.log(planet?.moons);       // 2
    console.log(planet?.diameter_km); // 6779

    // Or using response_format directly
    const rawResponse = await client.chat.completions.create({
      model: "openai/gpt-5.4",
      messages: [{ role: "user", content: "Tell me about Mars." }],
      response_format: {
        type: "json_schema",
        json_schema: {
          name: "planet_info",
          strict: true,
          schema: {
            type: "object",
            properties: {
              name: { type: "string" },
              diameter_km: { type: "number" },
              moons: { type: "integer" },
              habitable: { type: "boolean" },
            },
            required: ["name", "diameter_km", "moons", "habitable"],
            additionalProperties: false,
          },
        },
      },
    });

    const data = JSON.parse(rawResponse.choices[0].message.content ?? "{}");
    ```
  </Tab>

  <Tab title="cURL">
    ```bash theme={null}
    curl https://api.arouter.ai/v1/chat/completions \
      -H "Authorization: Bearer lr_live_xxxx" \
      -H "Content-Type: application/json" \
      -d '{
        "model": "openai/gpt-5.4",
        "messages": [
          {"role": "user", "content": "Tell me about Mars."}
        ],
        "response_format": {
          "type": "json_schema",
          "json_schema": {
            "name": "planet_info",
            "strict": true,
            "schema": {
              "type": "object",
              "properties": {
                "name": {"type": "string"},
                "diameter_km": {"type": "number"},
                "moons": {"type": "integer"},
                "habitable": {"type": "boolean"}
              },
              "required": ["name", "diameter_km", "moons", "habitable"],
              "additionalProperties": false
            }
          }
        }
      }'
    ```
  </Tab>
</Tabs>

## Model Support

`json_schema` mode with `strict: true` is supported by:

* `openai/gpt-5.4`, `openai/gpt-5.4-pro`, `openai/o3`, `openai/o4-mini`
* `anthropic/claude-sonnet-4.6`, `anthropic/claude-opus-4.5`
* `google/gemini-2.5-flash`, `google/gemini-2.5-pro`

`json_object` mode (no schema enforcement) is more broadly supported. Check `GET /v1/models` for the latest capability information.

## Streaming with Structured Outputs

Structured outputs work with streaming. The JSON content is delivered incrementally and you assemble it client-side:

```typescript theme={null}
const stream = await client.chat.completions.create({
  model: "openai/gpt-5.4",
  messages: [{ role: "user", content: "Tell me about Mars." }],
  response_format: {
    type: "json_schema",
    json_schema: {
      name: "planet_info",
      strict: true,
      schema: { /* ... */ },
    },
  },
  stream: true,
});

let jsonBuffer = "";
for await (const chunk of stream) {
  const content = chunk.choices[0]?.delta?.content;
  if (content) jsonBuffer += content;
}

const data = JSON.parse(jsonBuffer);
```

## Best Practices

1. **Use `strict: true`** — This guarantees schema compliance. Without it, the model may return valid JSON that doesn't match the schema exactly.

2. **Set `additionalProperties: false`** — Required for strict mode. Prevents the model from adding extra keys.

3. **List all required fields explicitly** — In strict mode, every field in `properties` should also be in `required`.

4. **Include a system prompt** — Telling the model its role as a data extraction or structured output assistant improves reliability.

## Error Handling

If the model cannot produce valid JSON that matches your schema (e.g., a prompt that fundamentally conflicts with the schema), the response will have `finish_reason: "length"` or `finish_reason: "content_filter"`. Always check `finish_reason` before parsing content.

```python theme={null}
response = client.chat.completions.create(...)

choice = response.choices[0]
if choice.finish_reason != "stop":
    raise ValueError(f"Unexpected finish reason: {choice.finish_reason}")

data = json.loads(choice.message.content)
```