Official OpenAI SDKs
Python, TypeScript and .NET clients work unchanged with a base URL and an API key.
Keep the SDKs, frameworks and tools you already use, and change one setting: the base URL. AI Server answers the familiar endpoints from models running on your machine, your network or your cluster.
| Endpoint | What it does |
|---|---|
POST /v1/chat/completions | Chat with streaming, tool calling for agents and image inputs for vision |
POST /v1/embeddings | Embeddings for search, retrieval and classification |
POST /v1/images/generations | Text-to-image, image-to-image and inpainting |
POST /v1/audio/speech | Text-to-speech with streamed audio |
POST /v1/audio/transcriptions | Speech-to-text from an uploaded file |
/v1/audio/transcriptions/stream | Live speech-to-text over WebSocket |
GET /v1/models | The models available to serve |
from openai import OpenAI
client = OpenAI(base_url="http://your-server:8080/v1", api_key="<api-key>")
reply = client.chat.completions.create(
model="<model-id>",
messages=[{"role": "user", "content": "Summarise our leave policy."}],
)
print(reply.choices[0].message.content)Python, TypeScript and .NET clients work unchanged with a base URL and an API key.
LangChain, Semantic Kernel, Microsoft Agent Framework and LlamaIndex.
Any tool with an “OpenAI-compatible” or custom base URL setting.
They find the server automatically and need no configuration.
The server will not serve a network until at least one API key exists.
Calls from the server’s own computer need no key.
On Free, AI Suite apps are unlimited and other OpenAI-compatible clients are rate-limited; any Pro edition removes the limit.
No. The OpenAI-compatible API is the SDK surface; use the official or community OpenAI client libraries.
Yes. /v1/chat/completions accepts tools and tool_choice ("auto", "none", "required") and returns tool_calls, streamed when stream is on.
AI Server does not offer the stateful Responses API (/v1/responses). Use /v1/chat/completions instead.
Free on one computer: unlimited for the AI Suite apps and rate-limited for other clients. Pro Personal and Pro Commercial serve your network without the limit.