<h2 id="simply-advanced-minimax-h3-workflow-by-altoiddealer">Simply Advanced - MiniMax H3 Workflow by altoiddealer</h2>This workflow was built from the ground up by myself working daily over the course of 2-3 weeks, aiming to be well structured, expandable, and easy to use without sacrificing any functionality.<h3 id="new-in-v1.5:">New in v1.5:</h3>Seamless Continuation has been replaced with Seamless Extension<ol><li>Slices off X frames from your video input(s) to set Video/Audio guide(s).</li><li>MiniMax H3 generates Video/Audio which begins and/or ends with those sliced inputs.</li><li>The workflow stitches the result with the original video input(s) using configurable overlap blending strategies. Optional: enter the subgraph to enable saving the raw generation result.</li></ol>Seamless Extension modes:<ul><li>Seamless Continuation (generate from input video as start point)</li><li>Seamless Convergence (generate into input video as end point)</li><li>Can be used in combination (Generate between two input videos)</li></ul><h2 id="main-features:">Main features:</h2><ul><li>Supports up to: 6 Images, 2 Videos, 3 Audio.<ul><li>ANY can be set as reference, guide, or both. This is much more useful than the average user understands - I have not yet seen this support in any other workflow.</li></ul></li><li>Latent Upscale and/or 2-Stage Generation<ul><li>Intuitive settings/LoRA management for each stage.</li></ul></li><li>1-4 steps LLM Prompt Enhancement<ul><li>Subgraph contains a dashboard to generate / edit LLM's responses before they pass into the next stage.</li><li>Includes my personal 4-Step Ref2VA pipeline.<ul><li>Step 1 Prompt Enhancement/Expansion: (Optional Step). Allows the LLM to interpret the user's prompt more verbosely and in relation to the reference materials. Catching/editing discrepancies here will yield much better results in next steps.</li><li>Step 2: Reference Analysis: Creates a detailed mapping of the references in relation to the Expanded Prompt. The resulting Reference Map is included as part of the context in the next steps.</li><li>Step 3: Creative Director: Considers Expanded Prompt + Reference Analysis to create a Creative Blueprint.</li><li>Step 4: Prompt Compiler: Conforms the Reference Map + Creative Blueprint into the H3 prompt structure</li></ul></li></ul></li><li>Includes my personal 1-Step FL2VA Prompt Enhance<ul><li>The system prompt begins with base instructions, adding task specific instructions (T2VA / I2VA / FL2VA / L2VA)</li><li>The mode is detected automatically based on the state of Image 1 / Image 2.</li></ul></li><li>You, the user, can easily add your own LLM instructions into this system as selectable options. Take a shot at making your own multi-stage pipeline!</li><li>Seamless video extension option - <a rel="ugc" href="https://civarchive.com/models/2881362/minimax-seed-hunter-workflow-latent-upscaler-seamless-video-continuation-speedups-ref-mods">continuation idea taken from Seed Hunter workflow</a> - later expanded to include convergence/combination usage.</li><li>Force Audio option - <a target="_blank" rel="ugc" href="https://civarchive.com/models/2881362/minimax-seed-hunter-workflow-latent-upscaler-seamless-video-continuation-speedups-ref-mods">idea taken from Seed Hunter workflow</a></li><li>Easily expandable framework<ul><li>Easily add/manage speedups, LoRAs, etc</li></ul></li><li>Is the poster child for an amazing new node, <a target="_blank" rel="ugc" href="https://github.com/yolain/ComfyUI-Easy-Use/pull/1035">Pass or None</a>. Workflow developers, start using this!!</li></ul><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/707d6bd3-7ae2-4512-9b56-46a248ed83c1/original=true/707d6bd3-7ae2-4512-9b56-46a248ed83c1.jpeg" /><h3 id="comfyui-requirements:">ComfyUI Requirements:</h3><ul><li>Be updated.</li><li>MUST have comfyui-frontend version &lt; <a target="_blank" rel="ugc" href="https://github.com/Comfy-Org/ComfyUI_frontend/pull/14475">v1.50.5</a> (Fixed broken Custom Combo nodes in subgraphs)</li></ul><h3 id="required-custom-nodes:">Required Custom Nodes:</h3><ul><li><a target="_blank" rel="ugc" href="https://github.com/yolain/Comfyui-Easy-Use">ComfyUI-Easy-Use</a></li><li><a target="_blank" rel="ugc" href="https://github.com/kijai/ComfyUI-KJNodes">ComfyUI-KJNodes</a></li><li><a target="_blank" rel="ugc" href="https://github.com/altoiddealer/comfyui_essential-er">comfyui_essential-er</a> (My repo). Was optional, now Required for v1.5+ for its <a target="_blank" rel="ugc" href="https://github.com/altoiddealer/comfyui_essential-er#merge-image-batches-and-audio-lists">superior video/audio merging node</a></li></ul><h3 id="optional-custom-nodes:">Optional Custom Nodes:</h3><ul><li>Load Video (Upload) from <a target="_blank" rel="ugc" href="https://github.com/kosinkadink/ComfyUI-VideoHelperSuite">ComfyUI-VideoHelperSuite</a> Feel free to replace with another Load Video node.</li><li><a target="_blank" rel="ugc" href="https://github.com/LBH-123-AI/Comfyui_Minimax_h3_latent_Upscaler">MiniMax H3 Latent Upscaler</a></li><li><a target="_blank" rel="ugc" href="https://github.com/Luisacaotica/ComfyUI-MiniMaxH3Mod">MiniMax H3 RefMods</a></li><li>Datetime String from <a target="_blank" rel="ugc" href="https://github.com/jamesWalker55/comfyui-various">comfyui-various</a> - Useful for naming outputs. NOT REQUIRED</li><li><a target="_blank" rel="ugc" href="https://github.com/xmarre/ComfyUI-Spectrum-MiniMax-H3">ComfyUI-Spectrum-MiniMax-H3</a> - Speedup option. NOT REQUIRED</li></ul>

Simply Advanced - MiniMax H3

v1.5.4: Fixed force_audio not actually working. Fixed optional match image input being disconnected in &lt;Picture 2&gt;.v1.5.3: Reverted the chain sampling implementation from 1.5.2. Currently too experimental.v1.5.2:<ul><li>D'ya know the expression, "you have to break a few eggs to make an omelet"? This fixes an issue where the "Second Pass" toggle had an inverted effect <code>Disabled = Yes / Enabled = No</code>. FIXED</li><li>Added improved chain sampling logic, so you can optionally return leftover noise from a not fully denoised First Pass.</li></ul>v1.5.1: Fix error when not using video inputs. Sorry to any who snatched v1.5 right away and experienced this.v1.5:<a target="_blank" rel="ugc" href="https://github.com/altoiddealer/comfyui_essential-er">ComfyUI-essential-er</a> (my repo) is now required.Reason: Extremely useful video/audio merging node which KJNodes has failed to merge from both <a target="_blank" rel="ugc" href="https://github.com/kijai/ComfyUI-KJNodes/pull/699">a PR I had opened</a>, as well as another user's <a target="_blank" rel="ugc" href="https://github.com/kijai/ComfyUI-KJNodes/pull/691">still open PR</a>.In terms of this workflow, it turns <a target="_blank" rel="ugc" href="https://i.imgur.com/vOTuBqV.png">this</a> (v1.4) into <a target="_blank" rel="ugc" href="https://i.imgur.com/sne4CXI.png">this</a> (v1.5).New Feature:"Seamless Continuation" has been replaced with "Seamless Extension"Seamless Extension modes:<ul><li>Seamless Continuation (generate from your video onwards)</li><li>Seamless Convergence (generate into your source video)</li><li>Can be used in combination</li></ul>

Files with SHA256: 2a9223a22bc6c65db44ded82683abae612b33edafd7797d348421aad9dba7c9d