Phase 6: ComfyUI — Workflow Engine · 55 min · Python · ComfyUI · websockets

ComfyUI API Integration

Design visually, execute programmatically — the /prompt, /history, /view API is how ComfyUI becomes production infrastructure.

What you will learn

Use the /prompt endpoint: POST API-format workflow JSON to trigger execution
Use the /history endpoint: poll for completed workflow results
Use the /view endpoint: retrieve generated output images/videos
Track execution via WebSocket events: node-by-node completion progress

Objective

Turn an exploratory graph into a reproducible, parameterized, inspectable production workflow. In the paired studio lab, build a Python client that: (1) loads an API-format workflow JSON, (2) submits it to ComfyUI's /prompt endpoint, (3) tracks progress via WebSocket, (4) retrieves the output image via /view, (5) saves it locally. Include error handling and timeout management.

The Problem

A team designs a great image generation workflow in ComfyUI's visual editor. Now they need to:

Execute it from a Python backend (not the UI)
Track execution progress (which node is running)
Retrieve the generated image programmatically
Handle errors and retries

They don't know ComfyUI has a REST API and WebSocket for programmatic control. They try to screen-scrape the UI — that's not production-grade.

ComfyUI API integration is the core skill for ComfyUI Pipeline Engineer roles — it's tested in every ComfyUI-focused interview.

What you'll build
Use the /prompt endpoint to trigger workflow execution, /history to poll for results, /view to retrieve output images, and WebSocket to track node-by-node execution progress. Build a complete Python client for ComfyUI.

Unlock the full lesson

You've read the first 2 sections. The rest of this lesson covers The Concept, Build, Check Yourself, Key Terms & Next — plus a hands-on lab, quiz, and project artifact.

Create a free account — the opening phases of 29 of 35 courses are free, no credit card.