Seedance-Style ComfyUI SD 1.5 Workflow with IPAdapter + Pose Control for Stylized Character Animation
Budget: €1,500 – €3,000 EUR
I'm looking for an expert in ComfyUI (Stable Diffusion 1.5 only) to create an easy to use modular, offline-ready animation workflow, specifically for stylized characters (e.g. Asterix, explainer cutouts, cartoon looks of any style).
Must-have features:
- Built for SD 1.5 only (no SDXL)
- Input: Reference image + text prompt → consistent character
- Full support for pose-to-pose animation:
- Start frame
- End frame
- Optional breakdowns (manually insertable)
- Style must stay 100% consistent across frames
- Use of IPAdapter, ReferenceOnly, or similar for style locking
- Pose control via ControlNet (OpenPose / Depth / Normal)
- Interface with drag-and-drop image loader nodes
Optional: Integration with Blender pose export (OpenPose JSON)
The goal is to replicate or exceed the functionality of Seedance, but for 2D stylized animation — not realism, not 3D.
Prompt-based Animation Control:
The animation should be driven by a text prompt, similar to how Seedance works.
The reference image defines the visual style and character appearance, while the text prompt determines what action or motion should happen (e.g. "walks forward", "turns around", "jumps and waves", etc.).
This prompt must influence the motion interpolation, pose selection, and overall animation flow – even when using fixed keyframes or poses.
The workflow should therefore combine:
Style/image-based control (IPAdapter / ReferenceOnly)
Prompt-based motion generation (AnimateDiff or similar)
Pose guidance (ControlNet keyframes)
Output:
HD default render (e.g. 768x512 or 1024x576)
Optional 4K export via upscaling nodes (latent upscaler / ESRGAN)
Switch between:
- Fast preview mode
- Final high-res render mode
Deliverables:
- Clean, working .json ComfyUI workflow file
- Node layout must be labeled and color-grouped
- All nodes installable via ComfyUI Manager
- Instructions (Markdown or quick video) to operate the flow
Exclude if:
- You only work with SDXL or realism
- You use Runway, EbSynth, Deforum, or cloud tools
- You do not know how to lock style across frames
Key Requirements:
- Workflow should support pose control using images/photos, video clips, and 3D models, as well as a pose editor
- Focus on stylized cartoon animation
- Color palette must be vibrant and colorful
Ideal Skills and Experience:
- Strong background in 2D animation and familiarity with cartoon styles
- Experience with ComfyUI and Stable Diffusion
- Proficiency in pose control techniques and tools
- Ability to create vibrant and colorful animations
Please provide samples of similar work done.
Core Workflow Requirements:
Base model: Stable Diffusion 1.5 (.safetensors)
Uses AnimateDiff Evolved or similar motion system
Text prompt defines the motion (as in Seedance)
Reference image defines the visual style / character
Must combine prompt-based animation control with pose-to-pose editing
Style locking via IPAdapter or ReferenceOnly
Pose control via ControlNet (OpenPose, Depth, Normal)
Integration of:
KJNodes or equivalent for keyframe control
CLIPVision encoder for style embedding
Platform:
Windows 11 with RTX Ada GPU (23.8 GB VRAM)
Language: English
Must-have features:
- Built for SD 1.5 only (no SDXL)
- Input: Reference image + text prompt → consistent character
- Full support for pose-to-pose animation:
- Start frame
- End frame
- Optional breakdowns (manually insertable)
- Style must stay 100% consistent across frames
- Use of IPAdapter, ReferenceOnly, or similar for style locking
- Pose control via ControlNet (OpenPose / Depth / Normal)
- Interface with drag-and-drop image loader nodes
Optional: Integration with Blender pose export (OpenPose JSON)
The goal is to replicate or exceed the functionality of Seedance, but for 2D stylized animation — not realism, not 3D.
Prompt-based Animation Control:
The animation should be driven by a text prompt, similar to how Seedance works.
The reference image defines the visual style and character appearance, while the text prompt determines what action or motion should happen (e.g. "walks forward", "turns around", "jumps and waves", etc.).
This prompt must influence the motion interpolation, pose selection, and overall animation flow – even when using fixed keyframes or poses.
The workflow should therefore combine:
Style/image-based control (IPAdapter / ReferenceOnly)
Prompt-based motion generation (AnimateDiff or similar)
Pose guidance (ControlNet keyframes)
Output:
HD default render (e.g. 768x512 or 1024x576)
Optional 4K export via upscaling nodes (latent upscaler / ESRGAN)
Switch between:
- Fast preview mode
- Final high-res render mode
Deliverables:
- Clean, working .json ComfyUI workflow file
- Node layout must be labeled and color-grouped
- All nodes installable via ComfyUI Manager
- Instructions (Markdown or quick video) to operate the flow
Exclude if:
- You only work with SDXL or realism
- You use Runway, EbSynth, Deforum, or cloud tools
- You do not know how to lock style across frames
Key Requirements:
- Workflow should support pose control using images/photos, video clips, and 3D models, as well as a pose editor
- Focus on stylized cartoon animation
- Color palette must be vibrant and colorful
Ideal Skills and Experience:
- Strong background in 2D animation and familiarity with cartoon styles
- Experience with ComfyUI and Stable Diffusion
- Proficiency in pose control techniques and tools
- Ability to create vibrant and colorful animations
Please provide samples of similar work done.
Core Workflow Requirements:
Base model: Stable Diffusion 1.5 (.safetensors)
Uses AnimateDiff Evolved or similar motion system
Text prompt defines the motion (as in Seedance)
Reference image defines the visual style / character
Must combine prompt-based animation control with pose-to-pose editing
Style locking via IPAdapter or ReferenceOnly
Pose control via ControlNet (OpenPose, Depth, Normal)
Integration of:
KJNodes or equivalent for keyframe control
CLIPVision encoder for style embedding
Platform:
Windows 11 with RTX Ada GPU (23.8 GB VRAM)
Language: English