Any model, any engine,
one local API.

from openai import OpenAI

client = OpenAI(base_url="http://127.0.0.1:11435/v1", api_key="unused-on-loopback")
reply = client.chat.completions.create(
    model="qwen3:8b",
    messages=[{"role": "user", "content": "Why is the sky blue?"}],
)
print(reply.choices[0].message.content)

curl http://127.0.0.1:11435/api/chat -d '{
  "model": "qwen3:8b",
  "messages": [{"role": "user", "content": "Hello"}]
}'

# any Ollama client: set OLLAMA_HOST=127.0.0.1:11435

import anthropic

c = anthropic.Anthropic(base_url="http://127.0.0.1:11435", api_key="alk_...")
m = c.messages.create(
    model="qwen3:8b", max_tokens=200,
    messages=[{"role": "user", "content": "Hello"}],
)

# deterministic answers with probabilities, scored from logits
curl http://127.0.0.1:11435/v1/decisions -d '{
  "model": "tev1:0.8b",
  "input": "I was charged twice for my order.",
  "questions": [
    {"type": "choice", "name": "route",
     "choices": [{"value": "billing"}, {"value": "shipping"}, {"value": "technical"}]},
    {"type": "predicate", "instructions": "Is the customer angry?"}
  ]}'

alkahest pull qwen3:8b                         # Ollama registry
alkahest pull hf.co/ggml-org/tiny-llamas:Q4_0  # Hugging Face
alkahest pull alkahest/qwen/qwen2.5-0.5b-instruct  # git catalog
alkahest import ~/models/qwen2-vl/ --link      # a folder, in place
alkahest serve                                 # 127.0.0.1:11435
alkahest ps                                    # what is loaded
alkahest keys create laptop --expires 30d      # share it