🗨️ Chat Completions

Method: chat()

Send chat completion requests to various LLM models through a unified interface.

📝 Example Usage

from omnilabs import OmniClient, ChatMessage

client = OmniClient()

# Using ChatMessage objects
messages = [
    ChatMessage(role="system", content="You are a helpful assistant."),
    ChatMessage(role="user", content="Tell me a joke.")
]

# Or using dictionaries
messages = [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "Tell me a joke."}
]

response = client.chat(
    messages=messages,
    model='gpt-4',
    temperature=0.7,
    max_tokens=100
)
print(response['content'])

🔹Parameters

Parameter	Type	Description
messages	List[Union[Dict[str, str], ChatMessage]]	List of messages. Each message can be a dictionary with 'role' and 'content' keys, or a ChatMessage object.
model	str	The model ID to use.
temperature	float	Sampling temperature (default: 0.7).
max_tokens	Optional[int]	Maximum tokens to generate. If None, uses model default.
stream	bool	Whether to stream the response (default: False).

🔹Response Format

{
    "model": "claude-3-5-sonnet-20241022",
    "content": "Generated content",
    "provider": "anthropic",
    "usage": {
        "all_tokens": 23,
        "total_cost": 0.00001
    }
}

---

Smart Model Selection

OmniRouter also provides a unique smart model selection feature through the smart_select() method:

response = client.smart_select(
    messages=[ChatMessage(role="user", content="Solve this calculus problem: ∫x²dx")],
    rel_accuracy=0.8,  # Prioritize accuracy
    rel_cost=0.2,
    verbose=True
)
print(f"Selected model: {response['model']}")
print(f"Response: {response['content']}")

This feature automatically selects the best model based on your specific needs, balancing factors like accuracy, cost, and latency.

---