​
Minimal Example

This snippet starts a client with OpenAI-compatible credentials and produces a single simulated conversation. Replace the placeholder keys with your own before running it in a script or notebook. 
import os
from collinear.client import Client

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

trait_basis = {
    "ages": ["25"],
    "genders": ["female"],
    "occupations": ["Employed"],
    "intents": ["check_flight_status"],
    "traits": {"patience": [1]},
    "locations": ["USA"],
    "languages": ["English"],
    "tasks": ["airline support"],
}

simulations = client.simulate(trait_basis, k=1, num_exchanges=2)
For richer scenarios, expand the trait_basis dictionary with additional persona buckets, intents, or turn counts before moving on to hosted evaluations.