One SDK. 300+ AI models.

OpenAI-compatible Python SDK for LLMs, Vision, OCR, Web Search, CDN, Short Links, and HTML Cleaning — unified in one interface.

$pip install sdkrouter

View Quickstart

quickstart.py

Python

from sdkrouter import SDKRouter

client = SDKRouter(api_key="YOUR_KEY")

res = client.chat.completions.create(
    model="openai/gpt-4o",
    messages=[{"role": "user", "content": "Hello!"}]
)

print(res.choices[0].message.content)

Loading code...

OpenAI-compatible API

Pydantic structured output

Async-ready

Cost tracking built-in

300+ models via OpenRouter

MIT License

Everything you need in one SDK

Unified interface for all your AI-powered features

Chat Completions

OpenAI-compatible chat interface with streaming support and async operations.

Structured Output

Native Pydantic integration for type-safe JSON responses and validation.

Vision Analysis

Image understanding with quality tiers: fast, balanced, and best accuracy.

OCR Extraction

Text extraction from images with multiple precision modes from tiny to maximum.

Web Search

Mode-based web search: search, research, analyze, comprehensive, investigate. With LLM ranking and entity extraction.

Short Links & CDN

URL shortening and CDN file uploads integrated into your workflow.

Why SDKRouter?

Built for developers who value speed, control, and reliability

Ship faster

One client, one interface. No juggling multiple SDKs or learning different APIs. Start building immediately.

Stay in control

Built-in models list, pricing data, and cost estimation. Know exactly what you're using and what it costs.

Build reliably

Timeouts, retries, async support, and typed responses. Production-ready features out of the box.

Pick speed vs accuracy

Quality tiers without rewriting code

Fast

Speed

gpt-4o-mini

Quick responses for high-volume tasks where speed matters most.

Balanced

Recommended

gpt-4o

Perfect balance of speed and accuracy for most use cases.

Best

Accuracy

claude-sonnet-4

Maximum quality for complex reasoning and detailed analysis.

Real-world examples

Copy, paste, and adapt to your needs

structured_output.py

Python

from pydantic import BaseModel, Field
from sdkrouter import SDKRouter

class Step(BaseModel):
    explanation: str = Field(description="Explanation of the step")
    result: str = Field(description="Result of this step")

class MathSolution(BaseModel):
    steps: list[Step] = Field(description="Solution steps")
    final_answer: float = Field(description="The final answer")

client = SDKRouter()
result = client.parse(
    model="openai/gpt-4o",
    messages=[
        {"role": "system", "content": "You are a math tutor."},
        {"role": "user", "content": "Solve: 3x + 7 = 22"},
    ],
    response_format=MathSolution,
)

solution = result.choices[0].message.parsed
print(f"Answer: x = {solution.final_answer}")

Loading code...

Built for production

Security and reliability you can trust

Secure by default

API keys handled securely with environment variable support. Never hardcode credentials in your code.

No vendor lock-in

OpenAI-compatible interface means you can switch providers without rewriting your application logic.

Self-hosted option

Optional custom base_url support for organizations that need to route through their own infrastructure.

Production-ready

Built-in timeouts, retries, error handling, and logging. Ready for high-scale production workloads.

Start routing AI in minutes.

Join developers building the next generation of AI applications with SDKRouter.

Get Started View on GitHub

$pip install sdkrouter