ComfyUI Cloud API

ComfyUI popularized node-based AI workflow design, but it runs locally and requires a dedicated GPU. Deploying it to the cloud — and wrapping it in a reliable API — typically means managing containers, GPU instances, cold starts, and custom endpoints. Most teams spend weeks on infrastructure before writing a single API call.

Wireflow takes a different approach. Workflows are built visually using a node canvas, and every published workflow is immediately accessible via POST /api/v1/workflows/{id}/execute. Authentication uses a standard bearer token. Execution is asynchronous — you get an executionId back, then poll GET /workflows/executions/{id}/poll until the job completes. No GPU procurement, no container orchestration, no custom server.

The API supports the same node types available in the UI: image generation (Flux 2 Pro, Recraft V4, Nano Banana 2), video generation (Kling, Veo3), post-processing (upscaling, background removal), and LLM-based prompt expansion. You send inputs, the cloud executes the graph, and you receive outputs — exactly the workflow pattern ComfyUI users already know, without the infrastructure burden.