Amazon SageMaker API: Full OpenAI-Compatible Support

The Amazon SageMaker API is now OpenAI-compatible: SageMaker AI real-time endpoints work with familiar tools — the OpenAI SDK, LangChain, Strands Agents — simply by changing the endpoint URL.

How It Works

Previously, integrating with SageMaker required custom clients and SigV4 signatures for AWS. This added layers of complexity and meant that developers either had to learn AWS specifics or write wrappers around standard tools. Now that complexity is a thing of the past.

If you're already using OpenAI API in your code, you can specify a SageMaker endpoint instead of OpenAI—and everything just works. No rewriting logic, no new dependencies in package.json or requirements.txt.

From your application's perspective, SageMaker simply "masquerades" as an OpenAI API. This is possible because OpenAI API has become the de facto standard in the ML community. When major providers support such an interface, engineers get genuine flexibility: they can switch between cloud platforms and choose the best option for each task without being locked into a specific API.

Concrete Example

Imagine you have a Python application using OpenAI. Previously, to switch to SageMaker, you would have had to rewrite much of the code. Now, just one line is enough:

client = OpenAI( api_key="unused", base_url="https://your-sagemaker-endpoint-url" )

The rest of the code remains completely unchanged. This works for LangChain, Strands Agents, and any other frameworks built on OpenAI API.

Who This Changes Life For

The change is useful for several groups of developers:

• Teams using LangChain—SageMaker is now a full-fledged model choice on par with OpenAI
• Those building multi-cloud systems—easier to work with multiple providers simultaneously
• AWS teams—embedding SageMaker models into existing code becomes trivial
• Startups avoiding lock-in—you can use AWS infrastructure without being tied to its API

This is especially valuable for companies that have already invested in the AWS ecosystem but want to remain flexible in choosing an LLM provider.

Why This Matters More Broadly

OpenAI API is the de facto standard in the industry. When major providers (AWS, Google, Meta) add compatibility, it means the LLM services market is becoming more competitive and mature. Developers no longer choose between competing integrations—they simply choose the best option based on price, latency, quality, and reliability. This is a signal: cloud platforms have realized that the era of lock-in is ending. The future belongs to hybridity, standardization, and freedom of choice.

What This Means

SageMaker becomes more competitive in the ML services market. For developers, this is a clean win: less boilerplate, more flexibility, easier to choose optimal solutions. For AWS, this is a step toward becoming a neutral platform for AI, not just a proprietary service. When API standards win, the entire industry wins.

*Meta is recognized as an extremist organization and banned in the Russian Federation.

Is SageMaker compatible with the OpenAI API?

Yes — Amazon SageMaker AI now supports an OpenAI-compatible API for real-time endpoints. Point your OpenAI client's base_url at your SageMaker endpoint URL (the article's example passes api_key="unused") and the rest of the code stays completely unchanged, including LangChain, Strands Agents, and other frameworks built on the OpenAI API. Previously this required custom clients and AWS SigV4 signatures.