Instructor
SDKs & frameworksStructured outputs as Pydantic models, with validation and retries.
Fill the guide with your values
Stays in this browser tab — never sent or stored by us. Sign in to create one
Setup
- 1Wrap an OpenAI client pointed at
https://api.ultragpt.pro/v1withinstructor.from_openai.
# pip install instructor
import os
import instructor
from openai import OpenAI
from pydantic import BaseModel
class User(BaseModel):
name: str
age: int
client = instructor.from_openai(OpenAI(base_url="https://api.ultragpt.pro/v1", api_key=os.environ["ULTRAGPT_API_KEY"]))
user = client.chat.completions.create(
model="grok-4.7",
response_model=User,
messages=[{"role": "user", "content": "John is 30 years old."}],
)Tips
- Any model id from Models works; suffixes like :online (web search) and :nitro work too.
- Give each tool its own key with a spend limit, so usage shows separately in Activity.
- Something off? Error codes · Support