What models does Qwen Multiangle need?

Qwen Multiangle requires four model files: qwen_image_edit_2511_bf16.safetensors (diffusion model), qwen_2.5_vl_7b_fp8_scaled.safetensors (text encoder), two LoRA files (multiangle and lightning), and qwen_image_vae.safetensors (VAE). Each goes in its respective ComfyUI/models subfolder.

Why are my Qwen Multiangle images not changing angles?

Check your camera angle settings in the Qwen Multiangle Camera node — an incorrect or unchanged angle value will produce similar results. Also make sure all model files are placed in the correct folders and that ComfyUI was restarted after installing the node.

How many angles can Qwen Multiangle generate?

Qwen Multiangle can generate up to 96 different camera angles from a single reference image, including front, back, side profiles, and various intermediate perspectives.

Qwen Multiangle in ComfyUI: Generate Images from Multiple Angles (2026)

Introduction

What this guide covers: How to install Qwen Multiangle in ComfyUI, download the required models, load the workflow, and generate consistent multi-angle images from a single reference — completely free.

Ever wanted to turn one image into many different camera angles while keeping everything else the same? With Qwen Multiangle and ComfyUI, you can do this for free — even as a beginner. One reference image in, up to 96 angle variations out.

What is Qwen Multiangle?

Qwen Multiangle is an open-source ComfyUI node that creates new images of the same scene or character from different camera angles. It takes your reference image and generates multiple views — side profiles, back shots, overhead angles, and more — while keeping the subject, details, and background consistent.

Single reference image

Just upload one image — Qwen Multiangle handles the rest.

Up to 96 angles

Generate front, back, sides, three-quarter views, and everything in between.

Consistent results

Character and environment details stay the same across every angle.

Examples: Generating Images from Multiple Angles

Here is what Qwen Multiangle produces from a single reference image. Each example below starts with one photo and generates four different camera-angle views while keeping the subject and environment identical.

Example 1 — Man Standing on a Rooftop

Reference Image

Generated Angles

Qwen Multiangle output — man on rooftop, angle 1: front view — Angle 1

Qwen Multiangle output — man on rooftop, angle 2: three-quarter view — Angle 2

Angle 3

Qwen Multiangle output — man on rooftop, angle 4: back view — Angle 4

Example 2 — Girl Character

Reference Image

Generated Angles

Qwen Multiangle output — girl character, angle 1: front view — Angle 1

Qwen Multiangle output — girl character, angle 2: three-quarter view — Angle 2

Angle 3

Qwen Multiangle output — girl character, angle 4: back view — Angle 4

Setting Up Qwen Multiangle in ComfyUI

The setup takes just a few minutes and only needs to be done once. Two parts: update ComfyUI, then install the Qwen Multiangle node.

Update ComfyUI First

Open ComfyUI Manager

Click the Manager button at the top of the ComfyUI interface.

Switch to the latest ComfyUI version

Click "Switch ComfyUI" and select the latest stable version.

Click "Update All"

This updates all installed nodes to their latest versions, which prevents compatibility issues with the new node.

ComfyUI Manager showing the Update All button to update all custom nodes — Click "Update All" in ComfyUI Manager before installing the Qwen Multiangle node

Install the Qwen Multiangle Node

Navigate to the custom_nodes folder

Open your ComfyUI installation folder and go into the custom_nodes subfolder.

Open a terminal in that folder

On Windows: type cmd in the address bar and press Enter. On Mac: right-click and select "New Terminal at Folder".

Windows Explorer address bar showing cmd typed to open Command Prompt inside the ComfyUI custom_nodes folder — Open a terminal directly inside the custom_nodes folder

Clone the Qwen Multiangle node repository

Run the command below in the terminal:

git clone https://github.com/jtydhr88/ComfyUI-qwenmultiangle.git

Command Prompt showing git clone command downloading the ComfyUI-qwenmultiangle repository successfully — Cloning the Qwen Multiangle node into your custom_nodes folder

⚠️ Note: If you do not have Git installed, download it first from git-scm.com/install before running the clone command.

Restart ComfyUI

After cloning finishes, close and reopen ComfyUI. The Qwen Multiangle nodes will appear in the node search.

Download and Place the Required Models

Before you can run Qwen Multiangle, you need to download five model files and place each one in the correct folder inside your ComfyUI installation. The table below has direct download links and the exact destination folder for each file.

Type	File	Folder	Download
Diffusion Model	`qwen_image_edit_2511_bf16.safetensors`	`models/diffusion_models/`	⬇ Get
Text Encoder	`qwen_2.5_vl_7b_fp8_scaled.safetensors`	`models/text_encoders/`	⬇ Get
LoRA — Multiangle	`qwen-image-edit-2511-multiple-angles-lora.safetensors`	`models/loras/`	⬇ Get
LoRA — Lightning (4-step)	`Qwen-Image-Edit-2511-Lightning-4steps-V1.0-bf16.safetensors`	`models/loras/`	⬇ Get
VAE	`qwen_image_vae.safetensors`	`models/vae/`	⬇ Get

Folder Structure

Once all five files are downloaded, your ComfyUI models directory should look like this:

📂 ComfyUI/
└── 📂 models/
    ├── 📂 diffusion_models/
    │      └── qwen_image_edit_2511_bf16.safetensors
    ├── 📂 loras/
    │      ├── qwen-image-edit-2511-multiple-angles-lora.safetensors
    │      └── Qwen-Image-Edit-2511-Lightning-4steps-V1.0-bf16.safetensors
    ├── 📂 text_encoders/
    │      └── qwen_2.5_vl_7b_fp8_scaled.safetensors
    └── 📂 vae/
           └── qwen_image_vae.safetensors

💡 Tip: After placing the files, refresh your browser tab so ComfyUI picks them up in the model dropdowns.

Using the Qwen Multiangle Workflow

Download the ready-made workflow below, load it into ComfyUI, and you are ready to start generating multi-angle images.

Download and open the workflow

Download the JSON file above, then drag it into ComfyUI to load it automatically.

Select the correct models in the workflow

In the workflow, find each model node and select the files you downloaded. Match each node type to its corresponding file.

ComfyUI Qwen Multiangle workflow showing model selection dropdowns for diffusion model, text encoder, LoRA, and VAE — Select the correct model files in each node of the workflow

Add your reference image

In the Load Image node, click Choose File and upload the reference photo you want to generate angles from.

ComfyUI Load Image node showing a reference portrait image uploaded and ready for Qwen Multiangle processing — Upload your reference image in the Load Image node

Set the camera angle in the Qwen Multiangle Camera node

Use the camera angle node to choose which view you want — front, back, side, three-quarter, and more. Watch the demo below to see how it works.

The Qwen Multiangle Camera node — select your desired camera angle and run the workflow

Click "Queue Prompt" to generate

ComfyUI will process your reference image through the Qwen Multiangle pipeline and output the new angle view. Run it multiple times with different angle settings for a full 360° set.

Tips and Troubleshooting

Images are not changing angles

Fix: Check the camera angle value in the Qwen Multiangle Camera node. An unchanged or incorrectly set angle will produce nearly identical outputs. Try a significantly different angle number.

Model errors when running the workflow

Fix: Make sure all five model files are in their correct folders (diffusion_models, text_encoders, loras, vae). The folder structure must match exactly. Refresh your browser after placing the files.

Missing node errors after loading the workflow

Fix: Restart ComfyUI after installing the node. If red errors still appear, go to Manager → "Install Missing Custom Nodes" to automatically install any additional dependencies.

Poor or inconsistent angle results

Fix: Use high-quality, well-lit reference images with a clear subject. Blurry or cluttered references produce inconsistent results. Portrait-style images with a plain or simple background work best.

💡 Tip: Generate all angles from the same seed for the most consistent results. Change only the camera angle between runs, and keep all other settings identical.

Frequently Asked Questions

Qwen Multiangle is an open-source ComfyUI node that generates new images of the same scene or character from different camera angles. It takes a single reference image and produces up to 96 angle variations while keeping the subject and background consistent throughout.

Navigate to your ComfyUI/custom_nodes folder, open a terminal, and run: git clone https://github.com/jtydhr88/ComfyUI-qwenmultiangle.git. Then restart ComfyUI. You also need to download the five required model files and place them in the correct folders before the node will work.

Qwen Multiangle needs five files: qwen_image_edit_2511_bf16.safetensors (diffusion model in models/diffusion_models), qwen_2.5_vl_7b_fp8_scaled.safetensors (text encoder in models/text_encoders), two LoRA files in models/loras (multiangle and lightning), and qwen_image_vae.safetensors in models/vae.

Check your camera angle settings in the Qwen Multiangle Camera node — an incorrect or unchanged angle value will produce similar results. Also verify all model files are placed in the correct folders and that ComfyUI was fully restarted after installing the node.

Qwen Multiangle supports up to 96 different camera angles from a single reference image. This includes front, back, side profiles, three-quarter views, overhead angles, and many intermediate perspectives — making it useful for character turnarounds and 3D-style image sets.