Skip to main content

​
Introduction

The ComfyUI REST API allows you to programmatically interact with ComfyUI to queue prompts, monitor execution, and manage workflows. The API uses JSON for request and response payloads.

​
Base URL

The default base URL for the ComfyUI API is: 
http://127.0.0.1:8188
All API endpoints can be accessed with an optional /api prefix for easier routing: 
http://127.0.0.1:8188/api/{endpoint}
Both formats are supported (e.g., /prompt and /api/prompt).

​
Authentication

​
API Nodes Authentication

If your workflow contains API nodes, you can include a Comfy API key in the extra_data field of your request payload: 
{
  "prompt": { /* your workflow */ },
  "extra_data": {
    "api_key_comfy_org": "comfyui-87d01e28d..."
  }
}
Generate your API key at https://platform.comfy.org/login

​
Request Format

All POST requests should:
  • Use Content-Type: application/json header
  • Send data as JSON-encoded request body

​
Response Format

All successful responses return JSON with appropriate HTTP status codes:
  • 200 - Success
  • 400 - Bad request (invalid prompt or parameters)
  • 403 - Forbidden (security violation)
  • 404 - Resource not found

​
CORS Support

CORS can be enabled using the --enable-cors-header command-line argument when starting ComfyUI. When running locally without CORS enabled, the server includes Origin validation for security.

​
WebSocket API

In addition to the REST API, ComfyUI provides a WebSocket endpoint for real-time updates: 
ws://127.0.0.1:8188/ws?clientId={client_id}
The WebSocket connection receives:
  • Execution status updates
  • Preview images during generation
  • Queue status changes
  • Error notifications
Use the WebSocket API in combination with the REST API to monitor prompt execution in real-time. See the websockets_api_example.py for a complete example.

​
Rate Limiting

There are no built-in rate limits, but be mindful of:
  • Server processing capacity
  • Queue size limits
  • Maximum upload size (configurable via --max-upload-size argument)

​
Common Workflow

  1. Export your workflow: Use “File → Export (API)” in the ComfyUI interface
  2. Queue the prompt: POST to /prompt endpoint
  3. Monitor execution: Connect to WebSocket or poll /history/{prompt_id}
  4. Retrieve results: GET images from /view endpoint using history data

​
Example: Basic API Usage

import json
from urllib import request

# Queue a prompt
prompt = {
    "3": {
        "class_type": "KSampler",
        "inputs": {
            "seed": 12345,
            "steps": 20,
            # ... other parameters
        }
    },
    # ... other nodes
}

data = json.dumps({"prompt": prompt}).encode('utf-8')
req = request.Request("http://127.0.0.1:8188/prompt", data=data)
response = request.urlopen(req)
result = json.loads(response.read())

print(f"Queued prompt ID: {result['prompt_id']}")

​
Next Steps

Queue Prompts

Learn how to submit workflows for execution

View History

Retrieve execution results and outputs

Manage Queue

Control the execution queue

Build docs developers (and LLMs) love

Get started for freeTalk to us