neverliie-ai-sdk 0.1.3

Minimal AI SDK for OpenAI, Anthropic, Google, and Mistral - optimized for Nuitka compilation

NeverLiie AI SDK

A minimal, unified Python SDK for interacting with multiple AI/LLM providers. Optimized for Nuitka compilation.

Features

• Multi-provider support - OpenAI, Anthropic (Claude), Google (Gemini), Mistral
• Minimal dependencies - Only requires requests>=2.28.0
• Streaming responses - Real-time content with SSE support
• Tool calling - Function/tool calling across all providers
• OpenAI-compatible - Generic provider for any OpenAI API-compatible endpoint
• Type safety - Full TypedDict support for messages, tools, and responses
• Nuitka optimized - Designed for compiling to standalone executables

Installation

pip install neverliie-ai-sdk

Quick Start

from neverliie_ai_sdk import Mistral

client = Mistral(api_key="your-api-key")
response = client.chat(messages="Hello, world!")
print(response["choices"][0]["message"]["content"])
client.close()

Supported Providers

Provider	Import	Default Model	Base URL
OpenAI	from neverliie_ai_sdk import OpenAI	gpt-4o-mini	https://api.openai.com/v1
Anthropic	from neverliie_ai_sdk import Anthropic	claude-3-haiku-20240307	https://api.anthropic.com/v1
Google	from neverliie_ai_sdk import Google	gemini-1.5-flash	https://generativelanguage.googleapis.com/v1beta
Mistral	from neverliie_ai_sdk import Mistral	mistral-small-latest	https://api.mistral.ai/v1
OpenAI Compatible	from neverliie_ai_sdk import OpenAICompatible	(configurable)	(configurable)

Usage Examples

Simple Chat

from neverliie_ai_sdk import OpenAI

client = OpenAI(api_key="your-api-key")

messages = [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "What's the capital of France?"}
]

response = client.chat(messages=messages, model="gpt-4o-mini")
print(response["choices"][0]["message"]["content"])

client.close()

Streaming Responses

from neverliie_ai_sdk import Anthropic

client = Anthropic(api_key="your-api-key")

for event in client.chat_stream(
    messages="Tell me a short story",
    model="claude-3-haiku-20240307"
):
    if event["type"] == "content":
        print(event["content"], end="")
    elif event["type"] == "tool_call":
        print("Tool call:", event["tool_call"])

client.close()

Tool Calling

from neverliie_ai_sdk import Google

client = Google(api_key="your-api-key")

tools = [{
    "type": "function",
    "function": {
        "name": "get_weather",
        "description": "Get the current weather for a location",
        "parameters": {
            "type": "object",
            "properties": {
                "location": {
                    "type": "string",
                    "description": "City and country"
                }
            },
            "required": ["location"]
        }
    }
}]

response = client.chat(
    messages="What's the weather in Paris?",
    model="gemini-1.5-flash",
    tools=tools,
    tool_choice="auto"
)

if response["choices"][0]["message"].get("tool_calls"):
    for tool_call in response["choices"][0]["message"]["tool_calls"]:
        print(f"Function: {tool_call['function']['name']}")
        print(f"Arguments: {tool_call['function']['arguments']}")

client.close()

OpenAI-Compatible Endpoints

from neverliie_ai_sdk import OpenAICompatible

# For OpenRouter, NVIDIA NIM, or any OpenAI-compatible API
client = OpenAICompatible(
    base_url="https://api.openrouter.com/api/v1",
    api_key="your-api-key"
)

response = client.chat(
    messages="Hello!",
    model="meta-llama/llama-3.1-8b-instruct"
)
print(response["choices"][0]["message"]["content"])

client.close()

API Reference

Client Initialization

All providers accept:

• api_key (str): Your API key for the service
• base_url (str, optional): Override the default base URL

Methods

chat(messages, model=None, tools=None, tool_choice=None, **kwargs)

Send a chat completion request.

Parameters:

• messages (str | list[dict]): User message string or list of message dicts
• model (str, optional): Model name (uses provider default if not specified)
• tools (list[dict], optional): List of tool definitions
• tool_choice (str, optional): Tool choice strategy ("auto", "required", or "none")

Returns: Response dict with normalized format

chat_stream(messages, model=None, tools=None, tool_choice=None, **kwargs)

Send a streaming chat completion request.

Returns: Iterator of event dicts with type field ("content" or "tool_calls")

close()

Close the HTTP session.

Error Handling

from neverliie_ai_sdk import Mistral
from neverliie_ai_sdk._exceptions import APIError, AuthenticationError, RateLimitError

client = Mistral(api_key="invalid-key")

try:
    response = client.chat(messages="Hello")
except AuthenticationError:
    print("Invalid API key")
except RateLimitError:
    print("Rate limit exceeded")
except APIError as e:
    print(f"API error: {e}")

client.close()

License

MIT License - see LICENSE file for details.