---
title: "vaklm - Python Client for OpenAI-Compatible APIs"
description: "A Python client for interacting with OpenAI-compatible API endpoints with reasoning support. Practical notes and code in JS."
url: "https://h3manth.com/scribe/vaklm/"
canonical_url: "https://h3manth.com/scribe/vaklm/"
markdown_url: "https://h3manth.com/scribe/blog/vaklm.md"
publishDate: "2025-01-26T13:28:47.050Z"
author: "Hemanth HM"
category: "python"
tags: ["python","ai","api","openai"]
word_count: 369
reading_time_minutes: 2
x_markdown_tokens: 886
content_signal: "ai-train=yes, search=yes, ai-input=yes"
---

# vaklm - Python Client for OpenAI-Compatible APIs

> A Python client for interacting with OpenAI-compatible API endpoints with reasoning support. Practical notes and code in JS.

vaklm was created to simplify intecctions with OpenAI-compatible APIs, providing a cleaner alternative to the verbose `client.chat.completions.create` pattern. This Python client offers a more intuitive interface while adding valuable reasoning capabilities to API responses.

vaklm is designed to make working with AI APIs more efficient and developer-friendly, providing an easy way to interact with OpenAI-compatible API endpoints while supporting reasoning capabilities.

## Installation
pip install vaklm

## Usage

### Basic Usage
from vaklm import vaklm, VAKLMException

try:
    content, reasoning = vaklm(
        endpoint="http://localhost:11434/v1/chat/completions",
        model_name="llama3.2:latest",
        user_prompt="Explain quantum computing in simple terms",
        system_prompt="You are a helpful AI assistant",
        api_key="YOUR_API_KEY",
        temperature=0.7
    )
    print("Content:", content)
    print("Reasoning:", reasoning)
except VAKLMException as e:
    print(f"Error: {str(e)}")

### Streaming Usagepython
from vaklm import vaklm

print("\nStreaming example:")
try:
    for content, reasoning in vaklm(
        endpoint="http://localhost:11434/v1/chat/completions",
        model_name="llama3.2:latest",
        user_prompt="Write a short story about a cat.",
        system_prompt="You are a creative writer.",
        api_key="YOUR_API_KEY",
        stream=True,
        temperature=0.7
    ):
        print(content, end='', flush=True)
        if reasoning:
            print(f"\n[Reasoning: {reasoning}]")
except VAKLMException as e:
    print(f"Error: {str(e)}")
```

## Features
- Supports both streaming and non-streaming responses
- Includes reasoning content in responses
- Automatic retry logic for failed requests
- Configurable temperature and max tokens
- System prompt support for context setting
- Comprehensive error handling

## Configuration
The `vaklm` function accepts the following parameters:

- `endpoint`: API endpoint URL (required)
- `model_name`: Model identifier (required)
- `user_prompt`: User's input message (required)
- `system_prompt`: Optional system context message
- `api_key`: API key for authentication
- `stream`: Whether to stream the response (default: False)
- `temperature`: Sampling temperature (0-2, default: 1.0)
- `max_tokens`: Maximum tokens to generate
- `timeout`: Request timeout in seconds (default: 30)
- `max_retries`: Maximum retry attempts (default: 3)
- `retry_delay`: Base delay between retries (default: 1.0)

## Error Handling
The client raises `VAKLMException` for general errors, with specific subclasses:
- `APIError`: For API-specific errors
- `StreamingError`: For streaming-specific errors

Always wrap calls in try/except blocks to handle potential errors.


`vaklm` provides a cleaner, more intuitive way to work with OpenAI-compatible APIs while adding valuable reasoning capabilities. By simplifying the API interaction pattern and adding features like automatic retries and streaming support, VAKLM helps developers focus on building great AI-powered applications rather than wrestling with API boilerplate. Give it a try in your next project and experience the difference!