Guides

The Guides section answers “how do I…” questions. Each page focuses on a single capability: it shows the minimal working code, explains the key parameters, and points out any platform-specific behaviour you need to know about. All examples use the OpenAI-compatible API, so existing OpenAI SDK code usually works with a one-line base URL change.

Pick a topic below to get started, or work through them in order if you are new to AI Foundation Services.

Core API

Chat Completions Send chat completion requests with system and user messages using the OpenAI-compatible Chat API on AI Foundation Services.

Streaming Stream LLM responses token-by-token via Server-Sent Events on AI Foundation Services for lower latency and better user experience.

Embeddings Generate text embeddings for semantic search, clustering, and RAG applications using the AI Foundation Services Embedding API.

Function Calling Define tools and let LLM Hub models return structured JSON arguments to integrate with external APIs and data sources.

Multimodal & Media

Multimodal (Vision) Analyze images alongside text using vision-capable models on AI Foundation Services with the standard Chat Completions API.

Audio (Speech-to-Text) Transcribe and translate audio files using Whisper-based models on AI Foundation Services, compatible with the OpenAI Audio API.

Image Generation Generate images from text prompts using the OpenAI-compatible Images API on AI Foundation Services with available image models.

Advanced

Reasoning Control Control chain-of-thought depth with the reasoning_effort parameter on supported models including o1, Gemini 2.5, and Claude 4.

Visual RAG Index documents with combined text and image retrieval on AI Foundation Services to extract data from charts, tables, and graphs.

Asynchronous Requests (Queue API) Submit long-running or batch inference jobs to the Queue API on AI Foundation Services and poll for results asynchronously.