Change a picture with up to five more pictures to take from. It is made for clothes: dress the person in your picture in what other pictures show, on the regular Qwen Image 2.1 model or the official Turbo checkpoint. It can also take a place, hair or light.Two versions on this page: v1.2 Regular runs the regular model at 25 steps, or 8 with a Fast switch that needs the Viggle Turbo LoRA. v1.2 runs the Turbo checkpoint in 8 steps. Same workflow, so pick the one that matches the model you have. The model list below is for Turbo; the Regular file's note has its own links.The same switches and upscale are in <a target="_blank" rel="ugc" href="https://youtu.be/ZX7RrPFGL_4">my Turbo Generate + Edit video on YouTube</a>.It's in my new video: <a target="_blank" rel="ugc" href="https://youtu.be/GociGWQQDSs">I threw a ton of edits at Qwen Image 2.1, FLUX 3 Klein and FIBO 2</a><h3 id="how-to-use-it">How to use it</h3><ol><li>Load the picture you want to change as Image 1.</li><li>Load the pictures to take from as Image 2 and up, and switch each one on in the card.</li><li>In the box over each picture, type what to take from it: "jacket", "red shorts", "outfit".</li><li>Run.</li></ol>It opens set up for clothes: Auto prompt is on clothes. The prompt box is optional: add anything else there, like "keep her own shoes". For another kind of edit, pick normal and say what to take, in the box over the picture ("hair") or in the prompt box: "Give the woman in image 1 the haircut of the woman in image 2."<h3 id="the-switches">The switches</h3><ul><li>Image 2 to Image 6. One switch for each picture to take from. Switch them on in order: a picture that is off is passed over and the next one takes its number.</li><li>Auto prompt: clothes, normal or off. Qwen looks at your pictures and writes the full instruction. It uses the text encoder the workflow already loads, so there is nothing more to download. What Qwen gets shows in Show Text.<ul><li>clothes: Qwen names each piece the boxes ask for and puts it into the sentence my Outfit Swap LoRA knows. An empty box takes that picture's outfit.</li><li>normal: anything that is not clothes. Qwen writes the edit from your prompt and the boxes, with my Consistency LoRA. "hair" over Image 2 takes the hair from image 2. An empty box adds nothing.</li><li>off: your prompt goes to Qwen as typed. The boxes are not read.</li></ul></li><li>Upscale the result. Redraws the result bigger in tiles with my Restore LoRA, to the Megapixels you set.</li></ul>Image 2 is on and Auto prompt is on clothes as the file opens. Image 3 to Image 6 and Upscale the result are off.Your own LoRAs go in the LoRA Loader, below my three rows. Mine are locked to their switches: Auto prompt turns on Outfit Swap for clothes or Consistency for normal and off, and Upscale the result turns on Restore. A strength you set stays, and your rows are never touched.<h3 id="what-you-need">What you need</h3><ul><li>AusBoss nodes 2.15.0 or newer (the locks on my LoRA rows). In ComfyUI-Manager, search "AusBoss". Already have them? Update first.</li><li>ComfyUI 0.38 or newer.</li><li>The models:<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/Qwen-Image-2.1/resolve/main/diffusion_models/qwen_image_2.1_turbo_int8_convrot.safetensors">qwen_image_2.1_turbo_int8_convrot.safetensors</a> → <code>models/diffusion_models</code> · 7.26 GB</li><li><a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/Qwen-Image-2.1/resolve/main/text_encoders/qwen3vl_8b_int8_convrot.safetensors">qwen3vl_8b_int8_convrot.safetensors</a> → <code>models/text_encoders</code> · 9.35 GB</li><li><a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/Qwen-Image-2.1/resolve/main/vae/qwen_image_2.1_vae_bf16.safetensors">qwen_image_2.1_vae_bf16.safetensors</a> → <code>models/vae</code> · 676 MB</li></ul></li><li>My Outfit Swap LoRA, for Auto prompt clothes: <a target="_blank" rel="ugc" href="https://huggingface.co/ausboss/Qwen-Image-2.1-Outfit-Swap-Consistency-LoRA/resolve/main/qwen-image-2.1-outfit-swap.safetensors">qwen-image-2.1-outfit-swap.safetensors</a> → <code>models/loras</code> · 159 MB</li><li>My Consistency LoRA, for normal and off: <a target="_blank" rel="ugc" href="https://huggingface.co/ausboss/Qwen-Image-2.1-Consistency-LoRA/resolve/main/qwen-image-2.1-consistency.safetensors">qwen-image-2.1-consistency.safetensors</a> → <code>models/loras</code> · 159 MB</li><li>Only for Upscale the result, my Restore LoRA: <a target="_blank" rel="ugc" href="https://huggingface.co/ausboss/Qwen-Image-2.1-Restore-LoRA/resolve/main/qwen-image-2.1-restore.safetensors">qwen-image-2.1-restore.safetensors</a> → <code>models/loras</code> · 159 MB</li></ul>You only need the one for the choice you use: a row that is off is not loaded.<h3 id="tips">Tips</h3><ul><li>The boxes do the naming. "The outfit from image 2" alone changed nothing in 7 of my 18 test swaps; with each piece named, all 18 went through. A word in a box is enough.</li><li>A wrong word gets drawn. If a piece comes out with the wrong color or pattern, read Show Text. Copy the sentence into your prompt, fix the word and run with Auto prompt off.</li><li>Clothes changed when you asked for something else? Pick normal, and take any clothes words out of the boxes.</li><li>With Auto prompt off, write a place as a change to Image 1: "Change the background of image 1 to the street from image 2." Asking to put her into image 2 gives back image 2.</li><li>A place comes through best from a picture without people in it. With people in it, normal leaves them out, but now and then the old background stays.</li><li>A place works best when Image 1 shows more than a face. With a close-up of a face and a picture of a whole street scene, Qwen gave back the street picture with her face on it.</li><li>Crop a picture to the thing you want taken from it. Drag the blue handles.</li><li>A piece from a picture of a whole outfit can bring the rest of that outfit along. Crop the picture to the piece, or use a picture of the piece alone.</li><li>One picture can hold several things. A jacket and a skirt side by side work as one picture.</li><li>Image 1 runs at 2 megapixels and the others at 1. Swaps came out best that way. To change it, set Megapixels on a picture node. For a bigger result, use the Upscale switch.</li><li>Shoes need the feet in Image 1. Leave shoes out when Image 1 does not show the feet: the result gets made-up legs or floating shoes.</li><li>Two people in Image 1: the clothes can go on the wrong one.</li><li>Hair, light and places: pick normal.</li><li>Times on an RTX 5090. About 10 to 15 seconds with two pictures, 25 with three, 35 to 40 with five or six. The upscale to 8 megapixels adds about 30.</li></ul><h3 id="what-it-does-not-do">What it does not do</h3><ul><li>A pose from another picture does not carry over. The person keeps the pose of Image 1.</li><li>It is not for swapping people. For a face or a whole person from another picture, use my <a target="_blank" rel="ugc" href="https://civarchive.com/models/2980685">Qwen Image 2.1 Person Swap</a>.</li><li>The upscale is a redraw. Fine detail is made up to fit, and small lettering can come back misspelled.</li><li>Everything was measured on one RTX 5090.</li></ul>For one outfit picture on the regular 25-step model, see my <a target="_blank" rel="ugc" href="https://civarchive.com/models/2983126">Qwen Image 2.1 Outfit Swap</a>. For new pictures and one-picture edits with the Turbo checkpoint, see <a target="_blank" rel="ugc" href="https://civarchive.com/models/2994247">Qwen Image 2.1 Turbo Generate + Edit</a>.<h3 id="credits">Credits</h3>Qwen Image 2.1 Turbo is by Alibaba Qwen, under the Qwen Research License: non-commercial use only. The ComfyUI files are from Comfy-Org.v1.2 (2026-10-20): Auto prompt picks clothes, normal or off; the boxes work with normal; one LoRA Loader with room for yours. v1.1 (2026-10-14): up to six pictures, a box over each picture. v1.0 (2026-10-13): first release.Questions or bugs: <a target="_blank" rel="ugc" href="https://github.com/ausboss/ComfyUI-AusBoss/issues">GitHub issues</a> or the comments here. Videos on <a target="_blank" rel="ugc" href="https://www.youtube.com/@ausbossdev">YouTube @ausbossdev</a>. More from me on <a target="_blank" rel="ugc" href="https://x.com/Zanzibased">X @Zanzibased</a> and <a target="_blank" rel="ugc" href="https://github.com/ausboss">GitHub</a>.

Qwen Image 2.1 Multi-Image Edit | Regular + Turbo

The same workflow on the regular Qwen Image 2.1 model. For people who don't have the Turbo checkpoint or want the full model. Everything is here: up to six pictures, the box over each picture, Auto prompt clothes, normal or off, the upscale.<ul><li>25 steps, with a Fast switch: the Viggle Turbo LoRA and 8 steps. It needs that LoRA, linked in the note.</li><li>About 25 seconds for a 2 megapixel edit with two pictures on an RTX 5090, about 8 with Fast.</li><li>The Turbo version stays the first tab on this page. Pick the one that matches the model you have.</li></ul>Needs ComfyUI 0.38 or newer and AusBoss nodes 2.15.0 or newer. Already have them? Update them in ComfyUI-Manager first, then restart ComfyUI.

Files with SHA256: f722debef78288650ef310ec20cc63e02e42c75a06d6b3e75bfc280a7e350173