Subconscious chat completions

Subconscious is OpenAI-SDK-compatible inference on a fine-tuned Qwen-3.6 served via sglang. You point an OpenAI client at our base URL; it works with standard OpenAI fields plus one knob for reasoning.

The whole product, in 6 lines

from openai import OpenAI

client = OpenAI(
    base_url="https://api.subconscious.dev/v1",
    api_key="sky_...",  # or set SUBCONSCIOUS_API_KEY
)

resp = client.chat.completions.create(
    model="subconscious/tim-qwen3.6-27b",
    messages=[{"role": "user", "content": "Hello"}],
)