Turn, crop and pad a picture, and Qwen Image 2.1 paints the new area. Straighten a crooked photo, cut away what you don't want, or open up any side. My outpaint LoRA keeps your picture in place, and only the new area and a thin strip along its edge change.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><h2 id="how-to-use-it">How to use it</h2><ol><li>Load your image. A new image resets the turn, crop and padding.</li><li>Turn it with the round handle, drag the crop edges in, and drag the orange handles out to add space.</li><li>Optional: finish the sentence in What goes there to say what belongs in the new area.</li><li>Press Run.</li></ol><h2 id="what-you-need">What you need</h2><ul><li>AusBoss nodes 2.5.0 or newer. In ComfyUI-Manager, search "AusBoss". Already have them? Update first.</li><li>A ComfyUI build with Text Encode Qwen Image 2.1 (mid-September 2026 or newer).</li><li>The Qwen Image 2.1 models and <a target="_blank" rel="ugc" href="https://civarchive.com/models/2967375">my outpaint LoRA</a>:<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_int8_convrot.safetensors">qwen_image_2.1_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><li><a target="_blank" rel="ugc" href="https://huggingface.co/ausboss/Qwen-Image-2.1-Outpaint-LoRA/resolve/main/qwen-image-2.1-outpaint-v2.safetensors">qwen-image-2.1-outpaint-v2.safetensors</a> → <code>models/loras</code> · 159 MB</li><li><a target="_blank" rel="ugc" href="https://huggingface.co/ausboss/Qwen-Image-2.1-Outpaint-LoRA/resolve/main/qwen-image-2.1-outpaint.safetensors">qwen-image-2.1-outpaint.safetensors</a> → <code>models/loras</code> · 159 MB (v1, for big zoom-outs)</li></ul></li></ul><h2 id="tips">Tips</h2><ul><li>Keep the LoRA on. Without it, plain Qwen 2.1 moved the scene in all 12 of my test renders on turned pictures, and the pasted-back picture showed doubled edges. With it, all 24 stayed put.</li><li>Finish the sentence in What goes there, for example <code>Also include in the prompt: a tall window on the left.</code> If you delete the first words, the workflow puts them back.</li><li>Pull out at most about your picture's own size on one side per run. More than that can paint a second scene. For more, run again on the result.</li><li>Crop out edges that confuse it. Noise, a frame line or a busy strip along the edge of your picture can make the new area come out strange. Drag the blue handles on Image Crop + Rotate + Pad inward to cut those edges off first.</li><li>Re-roll a part you don't like. Crop past it, then pad back out. That strip is painted fresh, and the prompt never sees what was there.</li><li>Big zoom-out? Switch the v2 row off and the v1 row on. v1 handles new areas much bigger than your picture.</li><li>About 8 seconds a run on an RTX 5090.</li></ul>Only need to extend a picture? Try <a target="_blank" rel="ugc" href="https://civarchive.com/models/2964810">Qwen Image 2.1 Outpaint</a>.<h2 id="credits">Credits</h2>Qwen Image 2.1 is © Alibaba Qwen, under the Qwen Research License: non-commercial use only, and commercial use needs a license from Qwen. The outpaint LoRA is mine and follows that license.v1.1 (2026-09-30): new layout, a blend-in seam, an empty image loader, and plain backdrops stay plain. v1.0 (2026-09-26): 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 Rotate + Outpaint

qwenImage21Rotate_v10.zip

Qwen Image 2.1 Rotate + Outpaint (AusBoss).json

v1.0

v1.1

<ul><li>The workflow is smaller: 13 nodes instead of 23, with the prompt and the painting in two boxes. Everything you change is on the top level.</li><li>The seam is set to blend in, a Stitch Inpaint option in AusBoss nodes 2.4.0. It fades the new area into your picture just inside the edge and leaves more of your picture untouched.</li><li>The image loader starts empty, so nothing shows up red when you open the workflow. The turn, crop and padding start at zero.</li><li>What goes there now starts with "Also include in the prompt:", so you only finish the sentence. Before, it held a line about the sample's bedroom.</li><li>Align is 1, so no gray strip gets added on the right and bottom anymore.</li><li>Your own words always reach the prompt. If you delete "Also include in the prompt:" and type your own line, the workflow puts those words back.</li><li>A plain backdrop stays plain. A product on a white studio sweep used to come back on a table; now the new area carries on the backdrop it already has.</li><li>Transparent PNGs work: the see-through parts are painted, and the prompt writer sees them as white, so a cutout gets a real backdrop.</li><li>Needs AusBoss nodes 2.5.0 or newer. The models and the LoRA are the same.</li></ul>

Files with SHA256: f68d1d4e13b25d124c95f68f2c3636d4e5766497c9a6acbf72fe4e49966d9018