https://youtu.be/949qxjckUyI 
 
FLUX 3 Klein 4K cloud batch upscale workflow (RunningHub online ZIP edition). Drop a ZIP of images on RunningHub and the whole folder is upscaled and re-generated at 4K with no local GPU. 
 
Main features: 
- HAIGC_LoadImagesFromZip: upload one ZIP (jpg/png), all images enter the queue automatically; HAIGC_SaveImagesToZip packs the 4K results back into a downloadable ZIP 
- Auto prompt reconstruction: RHLLMChatNode reads each image and writes the rebuild prompt (visual auto captioning, no manual prompting) 
- 16-step FLUX 3 Klein reconstruction: flux3_image_klein_sd_int8_convrot UNET, qwen3vl_4b CLIP, flux2-vae; KSampler 16 steps, CFG 1, euler/simple, fixed seed; ReferenceMultiLatent + ConditioningZeroOut keep the upscale faithful to the source 
- 4x AI super-resolution: RealESRGAN_x4plus (ImageUpscaleWithModel) after ImageScaleToMaxDimension 
- rgthree Image Comparer shows original vs 4K per image; SaveImage prefix AIKSK_FLUX3_KLEIN_FINAL_RH_BATCH 
 
Suggested workflow: 
1. Upload a ZIP of your images to RunningHub 
2. Queue the graph; unzip, caption, upscale, 16-step rebuild and re-zip all run automatically 
3. Download the result ZIP; compare 1K vs 4K in the comparer 
 
Dependencies: 
- flux3_image_klein_sd_int8_convrot.safetensors (FLUX 3 Klein UNET) 
- qwen3vl_4b_int8_convrot.safetensors (CLIP) 
- flux2-vae.safetensors (VAE) 
- RealESRGAN_x4plus.pth (upscale model) 
- HAIGC ZIP batch nodes, RHLLMChatNode, rgthree Node Suite 
 
RunningHub workflow page: https://www.runninghub.ai/zh-cn/post/2112077292083163137?inviteCode=rh-v1111 
Bilibili tutorial video: https://www.bilibili.com/video/BV11xJo6nEFh/ 
RunningHub registration with invite code rh-v1111 (1000 RH coins): https://www.runninghub.ai/?inviteCode=rh-v1111

FLUX 3 Klein 4K Cloud ZIP Batch Upscale

01_flux-3-klein-4k-cloud-zip-batch-upscale.json

v1.0

FLUX 3 Klein 4K Smart Upscale — RunningHub ZIP Batch

RunningHub workflow page: https://www.runninghub.ai/zh-cn/post/2112077292083163137?inviteCode=rh-v1111RunningHub China: https://www.runninghub.cn/post/2112055084015247361/?inviteCode=rh-v1111Bilibili tutorial video: https://www.bilibili.com/video/BV11xJo6nEFh/# FLUX 3 Klein — RunningHub ZIP Reconstruction Workflow## What this workflow doesStudy how a ZIP input and naming metadata are connected to individual-image and archive outputs.## Who should use itComfyUI users able to resolve compatible node versions, provide authorized images, and inspect actual outputs. This is one independent workflow, not a verified result package or a new model checkpoint.## RequirementsCommon loader references: `flux3_image_klein_sd_int8_convrot.safetensors`, `qwen3vl_4b_int8_convrot.safetensors`, `flux2-vae.safetensors`, and `RealESRGAN_x4plus.pth`. Weights, example inputs, and a tested installation lockfile are not included.List mapping and automatic per-image processing depend on installed node versions. A metadata connection is not proof of preserved names or complete archives. No batch run or cost measurement is included.## Configuration and stepsHAIGC_LoadImagesFromZip 17 feeds the upscaler, RH caption node 16, and comparer. Decoder 6 feeds SaveImage 5 and HAIGC_SaveImagesToZip 18. Naming information connects from loader output slot 6 to the ZIP saver’s naming input.Resolve the HAIGC plugin and service configuration first. The source note recommends testing three images of mixed orientations. Use distinguishable names and contents; check output count, name-to-image pairing, dimensions, individual saves, and archive completeness before a large batch. Node 5 uses prefix AIKSK_FLUX3_KLEIN_FINAL_RH_BATCH.The image path passes through model upscaling, ImageScaleToMaxDimension, ReferenceMultiLatent, sampling, and VAE decoding. The scaling node stores ["lanczos", 3840], and its title/note describe a long-edge target. ReferenceMultiLatent stores keep resolution. The sampler array is [9527, "fixed", 16, 1, "euler", "simple", 1]; the source titles/notes call it a 16-step/CFG1 preset, but the export lacks a complete named widget schema. Confirm the matching UI instead of inventing parameter mappings. No fixed 3840×2160 size, unchanged aspect ratio, improved quality, or successful output is verified.## Files- [Workflow JSON]()- [Promotional cover]()## Version note and limitationsThe JSON retains the supplied source bytes. This copy does not announce a new publication, model version, performance upgrade, or relationship to an existing remote project. Cover artwork is not a verified inference sample; supply a real reviewed sample separately if required.No new runtime, GPU, cost, quality, or batch-completion evidence is included. A requested caption, stored target, or progress value is not an output guarantee. Verify installation, image-transfer permissions, effective naming/type settings, and actual results before deployment.## Credits, tutorial, and rightsThe workflow is supplied through AI-KSK with source notes retained. A matching international tutorial URL has not been confirmed from the supplied metadata; no replacement is guessed. Verify rights for workflows, node code, weights, cover artwork, and source images separately. Regeneration does not remove source-image obligations or create blanket redistribution permission.

Files with SHA256: ba089a667213d02a5263683c7108b34d50f4b46c751ac6188d80116e301e9299