### 🎬 Roxy Potter: The Complete Video Comic Engine v3.5 (Base Edition)
An advanced, highly optimized ComfyUI workflow designed for generating sequential multimedia video comics without crashing your system.
#### 🚀 Key Features & VRAM Breakdowns:
* 8GB VRAM Friendly: Specifically tailored and re-written to run smoothly on budget and mid-range setups like the NVIDIA GeForce RTX 5050 8GB without facing Out-Of-Memory (OOM) errors.
* Hybrid Pipeline: Built to perfectly integrate local setup generation with cloud-based assets. It establishes a static starting frame using Krea 2 (Enhance/Cinematic) and passes it directly into the MiniMax (Hailuo AI) video generator node.
* Smart Scaling & Text Rendering: Features the custom DaSiWa Torch Resize node (Lanczos interpolation + active gamma correction) to upscale your final frames to 1920x1080 (Fit mode, scale_multiplier: 2.00) right before the video node. This keeps your embedded comic text overlays, background environments, and character assets crisp and readable without pixel degradation.
* 192-Frame Pacing (26 FPS): Configured for seamless 7.3-second cinematic shots with structured multimodal prompt layers including dedicated fields for description, character dialogue, soundscapes, and non-diegetic tracks.
#### 📦 How to Use:
1. Load your baseline character reference into the global identity node <Picture 1> is utilized as a global identity and stylistic visual reference).
2. Input your cinematic scene description and structured text into the Prompt Node.
3. Run the generation to output synchronized, full-HD video blocks complete with localized audio timing windows.
Brought to you by dim2. Free to download, explore, and run locally!
Description
### ⚡ Integrated Text-to-Prompt-to-Video Engine with High-FPS Interpolation (RTX 5050 8GB)
Initial release of the standalone, end-to-end local video comic engine pipeline. Custom-built to maximize generation speed, asset cohesion, and video smoothness within tight 8GB VRAM configuration ceilings.
#### 🔄 Complete End-to-End Workflow Architecture:
1. Stage 1: Identity Anchor & Text Processing
* Processes the global character style sheet <Picture 1> is utilized as a global identity and stylistic visual reference).
* Structuring text nodes using advanced multimodal formatting layouts split into description, raw dialogue, and multi-channel sound design arrays.
2. Stage 2: Hybrid Scaling Interface
Runs through the optimized *DaSiWa Torch Resize** integration loop utilizing standard Lanczos interpolation blocks and active gamma correction gamma_correct: true).
Targets a strict *192-frame pacing standard** locked natively to a smooth 26 FPS rendering profile.
3. Stage 3: Advanced MiniMax H3 Real-Time Rendering
* Feeds the upscale full-HD canvas into the MiniMax architecture, processing high-speed visual-text loops simultaneously to output smooth 7.3-second video clips without memory locks or application crashes.
No external LoRA dependencies required. Pure standalone optimization.
Looks like we don't have an active mirror for this file right now.
CivArchive is a community-maintained index — we catalog mirrors that volunteers upload to HuggingFace, torrents, and other public hosts. Looks like no one has uploaded a copy of this file yet.
Some files do get recovered over time through contributions. If you're looking for this one, feel free to ask in Discord, or help preserve it if you have a copy.