Learn / What is an OpenAI-compatible inference API?
What is an OpenAI-compatible inference API?
An OpenAI-compatible inference API is an API that follows the same request and response format as OpenAI's API, so developers can switch inference providers by changing the base URL and API key instead of rewriting their application. Netra Runtime provides an OpenAI-compatible API for running LLM and multimodal workloads in production.
Why it matters
OpenAI compatibility lets teams reuse existing SDKs, migrate workloads between providers, and drop a new backend into an existing app or agent framework without a rewrite. Because so much of the ecosystem already speaks the OpenAI format, compatibility is the most direct path to portability and the simplest way to avoid vendor lock-in.
What endpoints are commonly supported
| Endpoint | Purpose |
|---|---|
| Chat completions | Conversational and instruction-style generation |
| Completions | Plain text completion |
| Embeddings | Vector representations for search and RAG |
| Models | List the models available to serve |
| Streaming | Token-by-token responses for chat and agents |
How Netra uses it
Netra Runtime exposes an OpenAI-compatible API, so an application built with an OpenAI-style client can usually connect by changing the base URL and API key. That makes Netra a drop-in backend for chat apps, RAG systems, and agent frameworks already written against the OpenAI format. Learn what Netra Runtime is.
Frequently asked questions
What does OpenAI-compatible mean?
It means the API accepts the same requests and returns the same response shapes as OpenAI's API, so code and SDKs written for OpenAI work against it with only a base URL and API key change.
Can I use the OpenAI SDK with Netra Runtime?
Yes. Netra Runtime provides an OpenAI-compatible API, so applications using OpenAI-style SDKs can connect by changing the base URL and API key.
Do I need to rewrite my app to switch to Netra Runtime?
Usually not. If your app already uses an OpenAI-style client, you can point it at Netra Runtime by changing the base URL and key, which avoids most rewrites and reduces vendor lock-in.