Structured Outputs - MythicDot.AI

Why Structured Outputs?

✅

100% Compliance

Outputs always match your schema — no more parsing errors

⚡

No Post-Processing

Eliminate regex, retry logic, and validation code

🔧

Type Safety

Generate data that integrates directly with your code

Before vs After

❌ Without Structured Outputs

Hope the model returns valid JSON:

Parse errors at runtime
Missing required fields
Wrong data types
Retry logic needed
Complex validation code

✓ With Structured Outputs

Guaranteed valid JSON every time:

Schema-compliant output
All required fields present
Correct data types
No retry needed
Direct integration

Quick Start

Define your schema using JSON Schema and pass it to the API:

                    Python
                

from mythicdot import MythicDot

client = MythicDot()

response = client.chat.completions.create(
    model="mythic-4",
    messages=[
        {"role": "user", "content": "Extract info: John is 30, works at Acme Inc as an engineer"}
    ],
    response_format={
        "type": "json_schema",
        "json_schema": {
            "name": "person",
            "schema": {
                "type": "object",
                "properties": {
                    "name": {"type": "string"},
                    "age": {"type": "integer"},
                    "company": {"type": "string"},
                    "role": {"type": "string"}
                },
                "required": ["name", "age", "company", "role"]
            }
        }
    }
)

# Output is guaranteed to match schema
person = response.choices[0].message.parsed
print(person.name)  # "John"
print(person.age)   # 30
                

Schema Example

Person Schema

name string Full name of the person required

age integer Age in years required

company string Current employer required

role string Job title or position required

email string | null Email address if provided

Use Cases

📝 Data Extraction

Extract structured information from documents, emails, or any text into a consistent format.

🗃️ Database Population

Generate records that match your database schema exactly, ready for insertion.

🔄 API Responses

Build APIs that return consistent, typed responses from AI-generated content.

📊 Form Generation

Create form data, configuration files, or any structured content automatically.

Supported Types

Type	Description	Example
`string`	Text values	`"Hello"`
`integer`	Whole numbers	`42`
`number`	Floating-point	`3.14`
`boolean`	True/false	`true`
`array`	Ordered lists	`["a", "b"]`
`object`	Nested structures	`{"key": "value"}`
`enum`	Fixed set of values	`"draft" \| "published"`

💡 Pro Tip: Use Enums for Constrained Values

When you need the model to output only specific values (like status codes, categories, or ratings), use the enum type to guarantee valid outputs.

Start Building

Try structured outputs with your first API call.

Get Started →