Use the OpenAI SDK for images, REST for video async jobs, and webhooks for production pipelines.
The official @ai-sdk/openai provider gives you generateImage() with zero custom adapters.
Centralize routing with LiteLLM, generate typed SDKs from the OpenAPI spec, and use webhooks for async video.
Standard HTTP modules — no custom app or marketplace install needed. Works today.
Demo with cURL in 2 minutes. Hand the OpenAPI spec to any prospect for instant SDK generation.
All paths use the same api.creativeai.run/v1 endpoint and the same API key. Video generation is always async (submit + poll or webhook).
“I already use the OpenAI SDK”
Change api_key and base_url — done. All models work immediately.
“I am building a Next.js app”
Use the Vercel AI SDK with the official @ai-sdk/openai provider. generateImage() works out of the box.
“My team uses a proxy / gateway (LiteLLM)”
Hit the auto-config endpoint to get a ready-made YAML block. Zero manual editing.
“I need no-code automation (n8n / Make / Zapier)”
Use the standard HTTP module — no marketplace app or custom connector needed.
“I need video generation in production”
Submit a job via POST, receive an HMAC-signed webhook when it completes. No polling loops.
“I need to generate SDKs for an unsupported language”
Download the OpenAPI 3.1 spec and feed it to any code generator (openapi-generator, Speakeasy, etc.).