OpenRouter Provider

An unofficial Python wrapper for the OpenRouter API that provides a simple, intuitive interface for interacting with multiple LLM models. OpenRouter Provider supports chat conversations, image processing, tool integration, streaming responses, and structured output generation.

Features

• Conversation Memory: Automatic chat history management with easy memory control
• Image Processing: Built-in image resizing and base64 encoding for multimodal interactions
• Tool Integration: Decorator-based function calling with automatic tool execution
• Streaming Support: Real-time response streaming for both sync and async operations
• Structured Output: JSON schema-based response formatting using Pydantic models
• Async Support: Full async/await support for non-blocking operations
• Provider Configuration: You can select your favorite provoder easily

Quick start

Installation

pip install openrouter-provider

Set API key

Please get your API key from OpenRouter and set it as environment valuable.

OPENROUTER_API_KEY="your-api-key-here"

Basic Usage

from openrouter import *
from dotenv import load_dotenv

load_dotenv()

client = OpenRouterClient(system_prompt="You are a friendly AI assistant.")
model = LLMModel(name="openai/gpt-5")
query = Message(text="Hello, how are you?")

reply = client.invoke(model=model, query=query)
print(reply.text)

Timeout and retries

from openrouter import *

client = OpenRouterClient(
    system_prompt="You are a friendly AI assistant.",
    timeout=30,
    max_retries=0,
)

reply = client.invoke(
    model=LLMModel(name="openai/gpt-5"),
    query=Message(text="Hello"),
    max_completion_tokens=1000,
    timeout=10,
    max_retries=2,
)
print(reply.text)

OpenAI-Compatible Endpoints

from openrouter import *

client = OpenRouterClient(
    system_prompt="You are a helpful assistant.",
    base_url="http://localhost:11434/v1",
)

model = LLMModel(name="llama3.1")
query = Message(text="Hello, how are you?")

reply = client.invoke(model=model, query=query)
print(reply.text)