<p><strong>ByteDance USO makes subject and style fusion simple and powerful.</strong></p><p><strong>Who it's for:</strong> creators who want this pipeline in ComfyUI without assembling nodes from scratch. <strong>Not for:</strong> one-click results with zero tuning — you still choose inputs, prompts, and settings.</p><h3 id="open-preloaded-workflow-on-runcomfy">Open preloaded workflow on RunComfy</h3><p><a target="_blank" rel="ugc" href="https://www.runcomfy.com/comfyui-workflows/bytedance-uso-in-comfyui-unified-style-and-subject-workflow?utm_source=civitai&amp;utm_medium=referral">Open preloaded workflow on RunComfy (browser)</a></p><p><strong>Why RunComfy first</strong><br />- <strong>Fewer missing-node surprises</strong> — run the graph in a managed environment before you mirror it locally.<br />- <strong>Quick GPU tryout</strong> — useful if your local VRAM or install time is the bottleneck.<br />- <strong>Matches the published JSON</strong> — the zip follows the same runnable workflow you can open on RunComfy.</p><p><strong>When downloading for local ComfyUI makes sense</strong> — you want full control over models on disk, batch scripting, or offline runs.</p><p><strong>How to use (local ComfyUI)</strong><br />1. Load inputs (images/video/audio) in the marked loader nodes.<br />2. Set prompts, resolution, and seeds; start with a short test run.<br />3. Export from the Save / Write nodes shown in the graph.</p><p><strong>Expectations</strong> — First run may pull <strong>large weights</strong>; cloud runs may require a <strong>free RunComfy account</strong>.</p><hr /><h3 id="overview">Overview</h3><p>With this workflow, you can achieve seamless creative generation by combining both subject and style control into a unified process. It allows you to place characters into different scenes while preserving their identity with high accuracy. You can also apply diverse artistic styles from reference images to bring unique visual effects to your work. The system offers a combined mode that merges both subject-driven and style-driven guidance into one setup. It is built to deliver sharp, consistent, and flexible results for creators who value output precision. This makes it ideal for designers, digital artists, and content creators seeking efficient control over their visuals. The workflow streamlines generation tasks so you can achieve professional-grade imagery without a complex setup.</p><p><strong>Important nodes:</strong></p><h2 id="key-nodes-in-comfyui-bytedance-uso-workflow">Key nodes in Comfyui ByteDance USO workflow</h2><h3 id="usostylereference-(56)"><code>USOStyleReference</code> (#56)</h3><p>Applies a style image to the current model stream using the USO projector and CLIP‑Vision features. Use one reference for a strong, coherent look or chain two for nuanced blends; the second reference refines the first. If the style dominates too much, try a single, cleaner reference or simplify its content.</p><h3 id="referencelatent-(44)"><code>ReferenceLatent</code> (#44)</h3><p>Injects the encoded subject latent into the conditioning path so ByteDance USO preserves identity. Works best with uncluttered identity photos that clearly show the character’s face or defining features. If identity slips, feed a more complete reference or reduce conflicting style cues.</p><h3 id="fluxkontextmultireferencelatentmethod-(41)"><code>FluxKontextMultiReferenceLatentMethod</code> (#41)</h3><p>Combines multiple reference signals within the FLUX context pathway. This is where subject and prompt context are balanced before sampling. If results feel over‑constrained, relax references; if they drift, strengthen subject imagery or simplify the prompt.</p><h3 id="fluxguidance-(35)"><code>FluxGuidance</code> (#35)</h3><p>Controls the strength of text guidance relative to reference signals. Lower values let subject/style lead; higher values enforce the prompt more strongly. Adjust when you see either prompt underfitting (raise guidance) or style/subject being overridden (lower guidance).</p><h3 id="imagescaletomaxdimension-(109)"><code>ImageScaleToMaxDimension</code> (#109)</h3><p>Prepares the identity image for stable feature extraction. Smaller max sizes favor broader composition; larger sizes help when the reference is a tight portrait and you need crisper identity cues. Tune based on whether your subject ref is full‑body or a headshot.</p><h3 id="easycache-(95)"><code>EasyCache</code> (#95)</h3><p>Speeds up inference by reusing intermediate states when changes are minor. Great for prompt tweaks and rapid iteration, but it can slightly reduce micro‑details. Disable it for final, highest‑quality renders.</p><h3 id="ksampler-(31)"><code>KSampler</code> (#31)</h3><p>Runs the diffusion steps and controls stochasticity via seed and sampler choice. Increase steps for more detail, or lock the seed to reproduce a look while changing references. If textures look noisy, try a different sampler or fewer steps with stronger style guidance.</p><h3 id="notes">Notes</h3><p><strong>ByteDance USO in ComfyUI | Unified Style and Subject Workflow</strong> — see <a target="_blank" rel="ugc" href="https://www.runcomfy.com/comfyui-workflows/bytedance-uso-in-comfyui-unified-style-and-subject-workflow?utm_source=civitai&amp;utm_medium=referral">RunComfy page</a> for the latest node requirements.</p>

ByteDance USO in ComfyUI | Unified Style and Subject Workflow

<p>Initial release — ByteDance-USO.</p>

Files with SHA256: 68c579de03bfaf6744640ad3ed2d003b81b5b56f5e9990f49dbb1168dcc8769b