​
1. Create an environment with uv

curl -LsSf https://astral.sh/uv/install.sh | sh
uv venv --python 3.12
source .venv/bin/activate
uv pip install --upgrade pip collinear

​
2. Create a short script quickstart.py

from collinear.client import Client

client = Client(
    assistant_model_url="https://api.openai.com/v1",
    assistant_model_api_key=OPENAI_API_KEY,
    assistant_model_name="gpt-4o-mini",
    steer_api_key=TRAIT_BASIS_API_KEY,
)

trait_basis = {
    "ages": ["13-17", "18-24", "25-34", "35-44", "45-54", "55-64", "65+"],
    "genders": ["male", "female", "other"],
    "occupations": ["Unemployed", "Employed", "Student", "Retired", "Not in Labor Force"],
    "intents": [
        "search_flights",
        "make_booking",
        "modify_booking",
        "cancel_booking_request_refund",
        "track_baggage",
    ],
    "traits": {
        "impatience": [0, 1, 2],
        "confusion": [0, 1, 2],
        "skeptical": [0, 1, 2],
    },
    "locations": ["USA", "Canada", "UK", "Australia", "other"],
    "languages": ["English", "Spanish", "French", "other"],
    "tasks": ["airline support"],
}

conversations = client.simulate(trait_basis, k=3, num_exchanges=3)
assessment = client.assess(conversations)
print(assessment)
The SDK validates these persona buckets. Supplying unsupported ages or occupations raises immediately, while combinations that pair "Retired" with an age bucket below "35-44" are simply skipped. Replace the placeholder keys with your own before running.

​
3. Execute the run

uv run python quickstart.py
The script leaves you with two Python objects:
  • conversations: the simulated Trait Basis conversations.
  • assessment: safety scores for each conversation.
Inspect them in your editor or an interactive shell, then move on to the Trait Basis simulation guide for deeper configuration or the Platform API when you need hosted runs.