Image, Audio, and Multimodal Routing
Same gateway, multiple modalities — vision models, Whisper-style audio, image generation. Proves the OpenAI-compatible API is not just a chat surface.
💡
Tip: Most teams stop at chat. The same gateway handles vision, audio, and image generation with the same governance, the same observability, and the same per-key controls.
What this use case covers
- Vision models (image-in, text-out) routed through the same proxy.
- Audio — Whisper-style transcription, ElevenLabs / Polly synthesis — with the same key, budget, and audit story.
- Image generation (DALL-E, Stable Diffusion endpoints) behind the gateway.
- A single client that talks to all four modalities through one base URL, with per-modality budgets and rate limits.