Extend a picture on any side, or turn it first, and FLUX 3 Image Klein paints the new area. A slider shows the picture before and after.The LoRA it uses: <a target="_blank" rel="ugc" href="https://civarchive.com/models/3003865">FLUX 3 Image Klein Outpaint LoRA</a>. Also for FLUX 3 Klein: <a target="_blank" rel="ugc" href="https://civarchive.com/models/3004026">Single &amp; Multi-Image Edit</a>.<h3 id="how-to-use-it">How to use it</h3><ol><li>Load your image.</li><li>Pick a ratio, or drag the orange handles out. To straighten a crooked picture, turn it with the round handle.</li><li>Optional: finish the sentence in What goes there.</li><li>Run.</li></ol>Your original pixels are put back on top at the end, so only the new area is painted.<h3 id="what-you-need">What you need</h3><ul><li>ComfyUI 0.41.0 or newer. That is the first version with FLUX 3.</li><li>AusBoss nodes 2.13.0 or newer. In ComfyUI-Manager, search "AusBoss". Already have them? Update first.</li><li>The FLUX 3 Image Klein files. They are on Hugging Face behind a license click: open <a target="_blank" rel="ugc" href="https://huggingface.co/black-forest-labs/flux-3-image-klein-comfy">black-forest-labs/flux-3-image-klein-comfy</a>, sign in and accept, then these links work.<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/black-forest-labs/flux-3-image-klein-comfy/resolve/main/diffusion_models/flux3_image_klein_sd_int8_convrot.safetensors">flux3_image_klein_sd_int8_convrot.safetensors</a> → <code>models/diffusion_models</code> · 5.83 GB</li><li><a target="_blank" rel="ugc" href="https://huggingface.co/black-forest-labs/flux-3-image-klein-comfy/resolve/main/text_encoders/qwen3vl_4b_int8_convrot.safetensors">qwen3vl_4b_int8_convrot.safetensors</a> → <code>models/text_encoders</code> · 4.86 GB</li><li><a target="_blank" rel="ugc" href="https://huggingface.co/black-forest-labs/flux-3-image-klein-comfy/resolve/main/vae/flux2-vae.safetensors">flux2-vae.safetensors</a> → <code>models/vae</code> · 336 MB</li></ul></li><li>My outpaint LoRA: <a target="_blank" rel="ugc" href="https://huggingface.co/ausboss/FLUX-3-Image-Klein-Outpaint-LoRA/resolve/main/flux-3-image-klein-outpaint.safetensors">flux-3-image-klein-outpaint.safetensors</a> → <code>models/loras</code> · 161 MB. If the run stops at a red LoRA Loader, this file is missing.</li></ul>These are the INT8 files, the same ones ComfyUI's own FLUX 3 template uses. Black Forest Labs says they run on cards with 8 GB. The full-size BF16 files are on the same page for bigger cards (11.1 GB and 8.88 GB): pick them in Diffusion model, Text encoder and Caption model.<h3 id="tips">Tips</h3><ul><li>Pull out about your picture's own size at most, on one side per run. For more, run again on the result.</li><li>Leave the fill and the feather alone. The LoRA learned a hard gray <code>#808080</code> edge, so the Fill stays <code>#808080</code> and Feather stays 0.</li><li>Not happy with the new area? Run it again. The seed changes each run. In my trial runs, one in 18 drew part of the picture a second time, and at 1.5 one in 20.</li><li>Say what should be there only when you need to. What goes there tells the text encoder what else to put in the description of your picture.</li><li>About 8 seconds a run on an RTX 5090.</li></ul>The LoRA is what keeps your picture in place. Without it FLUX 3 often zooms in or slides the picture instead of extending it: 36 of 40 runs in my test of the official template graph. The LoRA, its test numbers and its license are on the <a target="_blank" rel="ugc" href="https://huggingface.co/ausboss/FLUX-3-Image-Klein-Outpaint-LoRA">Hugging Face repo</a>.On Qwen Image 2.1 instead: <a target="_blank" rel="ugc" href="https://civarchive.com/models/2964810">Qwen Image 2.1 Outpaint</a> and <a target="_blank" rel="ugc" href="https://civarchive.com/models/2966883">Qwen Image 2.1 Rotate + Outpaint</a>.<h3 id="credits">Credits</h3>FLUX 3 Image Klein is by Black Forest Labs, under the FLUX Kommunity License: use of the pictures you make is covered by its rules for outputs (section 2), so read it if you plan to use them commercially. The outpaint LoRA is mine, a modified version of FLUX 3 Image Klein shared under the same license. The files are Black Forest Labs' own ComfyUI versions.v1.0 (2026-10-15): 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>.

FLUX 3 Image Klein Outpaint

First release. Needs ComfyUI 0.41.0 or newer and AusBoss nodes 2.13.0 or newer.

Files with SHA256: 705703e1bb461c1fa32124db3bd118706478c97b21a6f10112f2e426d6aea156