HACKAIGCHACKAIGC

ComfyUI + OpenRouter Guide: How to Access 20+ LLMs in One Workflow

Elizabeth Rowan Carteron 11 hours ago

Introduction

ComfyUI is the most popular node-based AI image generation workflow tool, giving creators granular control over every step of the generation process. But it has always had a limitation: you're basically locked into whatever model you load into your workflow.

Want to switch from GPT to Claude mid-workflow? Previously, you'd have to export, switch tools, and re-import. It was clunky.

In late May 2026, that changed. OpenRouter officially announced native ComfyUI support, allowing you to call 20+ models directly from within your ComfyUI workflow.

This guide walks you through the setup, the use cases, and why this integration is a game-changer for uncensored AI creators.

What Is ComfyUI + OpenRouter?

ComfyUI

ComfyUI is a node-based interface for AI image generation. You connect visual nodes — prompt input, model loader, sampler, post-processing — to build custom generation pipelines. It's the preferred tool for power users who want full control.

OpenRouter

OpenRouter is an AI model aggregator. One API key gives you access to GPT-5.5, Claude Opus 4.7, Gemini 3.1 Pro, Llama 4, DeepSeek V4, and 20+ other models. You pay per token, no separate accounts needed for each model.

Why They Work Together

The integration decouples the model selection layer from your workflow. Instead of rebuilding your node graph to switch models, you just change a dropdown on the OpenRouter node.

What This Integration Brings

According to the official announcement from OpenRouter and ComfyUI:

  1. Native OpenRouter node — drop it into any workflow

  2. 20+ models available — both commercial and open-source

  3. Unified billing — one OpenRouter account, one API key

  4. Community workflows — official example workflows already published

Step-by-Step Setup Guide

Prerequisites

  • ComfyUI installed (latest version recommended)

  • OpenRouter account

  • OpenRouter API Key

Step 1: Get Your OpenRouter API Key

  1. Go to OpenRouter.ai and log in

  2. Navigate to Settings → API Keys

  3. Click Create Key, copy the key

  4. Save it somewhere secure (it won't be shown again)

Step 2: Install the OpenRouter Node in ComfyUI

  1. Launch ComfyUI

  2. Open the Manager → Install Custom Nodes

  3. Search for OpenRouter or the relevant integration node

  4. Install and restart ComfyUI

Step 3: Configure the Node

  1. Drag an OpenRouter LLM node into your workspace

  2. Paste your API key into the node's configuration

  3. Select your target model from the dropdown (e.g. GPT-5.5, Claude Opus 4.7)

  4. Set parameters like temperature and max_tokens

Step 4: Connect to Your Workflow

  1. Route the OpenRouter LLM node output to your image generation node's prompt input

  2. Configure your generation parameters

  3. Run a test

Step 5: Switch Models On the Fly

Once it works, try switching models. Go back to the OpenRouter node, pick a different model, and run again. No other part of your workflow needs to change.

Why This Matters for Uncensored Creators

If you generate NSFW or uncensored content, this integration is especially valuable:

1. Bypass Model-Level Restrictions

Mainstream models like GPT and Claude have strict content filters. With ComfyUI + OpenRouter, you can switch to a more permissive model (like DeepSeek V4 or a community fine-tune of Llama 4) — without switching tools.

2. Compare Models Side-by-Side

Same workflow, same prompt, different models. See which one gives the best results for your specific use case, directly inside ComfyUI.

3. Lower Experimentation Cost

OpenRouter charges per token. You can test a dozen models for a few cents before committing to one.

When You Just Want to Create

Of course, if you don't want to set up nodes, configure APIs, or manage model switching at all — HackAIGC offers a plug-and-play experience for uncensored image, video, and chat generation. No setup, no filters.

Practical Use Cases

Use Case 1: Multi-Model Prompt Engineering

Use GPT-5.5 to draft a prompt, then Claude Opus 4.7 to refine the style description. Both inside the same ComfyUI canvas.

Use Case 2: Uncensored Generation

Run your diffusion model as usual, but use a permissive LLM (like DeepSeek V4) through OpenRouter for prompt enhancement. The LLM handles the text, your image model handles the visuals.

Use Case 3: Workflow Testing

Test multiple models on the same prompt to find the best fit for your niche. Log each result, compare outputs, pick your winner.

ComfyUI + OpenRouter vs Alternatives

Approach

Models Available

Setup Complexity

Content Filters

ComfyUI + OpenRouter

20+

Low

Depends on model chosen

ComfyUI local model

1–3

High (download weights)

None (local)

Direct API in web UI

1

Medium

Depends on provider

HackAIGC (one-click)

Optimized engine

Zero

None (uncensored)

For flexibility and model choice, ComfyUI + OpenRouter leads. For zero-config uncensored creation, HackAIGC is simpler.

Frequently Asked Questions

Q: Is ComfyUI + OpenRouter free? A: ComfyUI is free and open-source. OpenRouter charges per token, starting at $0.15/million tokens.

Q: Can I use multiple models in one workflow? A: Yes. Just add multiple OpenRouter nodes, each configured with a different model.

Q: Do I need to rebuild my workflow when switching models? A: No. Only change the model selection on the OpenRouter node.

Q: Can I mix local models with OpenRouter models? A: Yes. Local model nodes and OpenRouter nodes coexist in the same workflow.

Q: I just want to generate uncensored content without any setup. What should I use? A: Try HackAIGC — it's designed for instant, uncensored generation with no configuration needed.

Try HackAIGC products: