OpenAI
GPT family LLM, embeddings, image / video generation, STT, TTS, vision
Vendor ID: openai · Maturity: preview · openai.com · Docs · Categories: AI
OpenAI AI models and APIs including GPT and DALL-E. The LLM speaks openai-chat-v1, which is also the canonical wire protocol for any OpenAI-compatible vendor.
Auth
| Credential | Notes |
|---|---|
apiKey | OpenAI API key. Sent as Authorization: Bearer <key>. |
Products
OpenAI (root product)
ID: openai:openai — exposes embeddings and the image-generator entry point at the vendor root.
Image Generator
ID: openai:image-generator — DALL-E and successor models for image generation. Powers the Image capability runtime.
Speech-to-Text
ID: openai:stt — Whisper-family transcription. Contributes the STT capability.
Text-to-Speech
ID: openai:tts — voice synthesis. Contributes the TTS capability.
Video Generator
ID: openai:video-generator — video generation. Contributes the Video capability.
Vision
ID: openai:vision — image analysis (description, OCR, structured extraction). Used by the Agent block when the selected model is vision-capable.
Capabilities
| Capability | Default model | Wire protocol |
|---|---|---|
| LLM | gpt-5.4 | openai-chat-v1 |
| Embedding | per embeddings product | — |
| Image | per image-generator product | — |
| STT | per stt product | — |
| TTS | per tts product | — |
| Video | per video-generator product | — |
| OCR / Vision | per vision product | — |
All capabilities resolve through the integration registry. The Agent / Router / Evaluator blocks never reference openai by name.