ComfyUI WANv2v Video Stitcher & ControlNet Manager

🚀 ComfyUI custom node pack for WAN v2v long-video generation.
This pack provides all the tools you need to build seamless long-form videos using chunking + overlap stitching techniques. It includes debug nodes, iteration-based control group orchestration, and a de-overlap merge node to produce smooth transitions without frame duplication.

---

✨ What You Get with This Pack

• Seamless long video generation: Split videos into overlapping chunks, process with WAN v2v, then stitch without artifacts.
• Chunk + overlap orchestration: Handle chunk_size and overlap_size automatically across iterations.
• De-overlap merging: Drop duplicate frames from overlapping regions for a clean final sequence.
• Debug-friendly workflow: Visualize frame numbering to ensure sequential order without skips or repeats.
• Ready for ControlNet pipelines: Works with ControlNet batches for consistent conditioning.

This pack is especially useful if you are generating videos with WAN v2v where the KSampler must process limited frame chunks (e.g., 81 frames) and you want to stitch them into one continuous video.

---

🧩 Nodes Included

1. WanV2V Debug: Number Batch

• Input: IMAGE batch
• Output: Same batch with numbers stamped in the top-right corner (1,2,3...).
• Purpose: Debug sequential frame flow across nodes.
• ✅ Helps verify frame order and detect overlaps/skips.

---

2. WanV2V: Iter Control Group

• Inputs:
  • controlnet_images (full ControlNet frame sequence)
  • prev_output (optional, previous sampler output)
  • chunk_size, overlap_size, iter, fps
• Outputs:
  • control_group (IMAGE batch for this iteration)
  • info (summary string with frame ranges, drops, padding info)
• Purpose: Builds each chunk for WAN v2v long-video generation.
• Logic:
  • Iter 1: Takes first chunk_size frames.
  • Iter ≥2: Prepends last x frames from previous sample, adds residue from prior group, and appends current frames while dropping overlaps.
  • Final Iter: Supports conclusive stitching by padding tail frames if ControlNet frames run out.
• ✅ Ensures no skipped or duplicated frames across iterations.

---

3. WanV2V: Merge De-Overlap (10x)

• Inputs: Up to 10 IMAGE batches (image_1 ... image_10), plus overlap_size.
• Output: A single merged sequence with overlaps removed.
• Purpose:
  • Removes first x frames from every batch except the first.
  • Concatenates sequentially → clean video-ready batch.
• ✅ Final step before video export.

---

🔧 Installation

1. Navigate to your ComfyUI custom_nodes folder:

   cd ComfyUI/custom_nodes
   
   

2. Clone this repo:

git clone https://github.com/<your-username>/comfyui-wanv2v-video-stitcher.git

3. Restart ComfyUI. The new nodes will appear under:

WanV2vControlnetManager/Debug

WanV2vControlnetManager/WAN

🚀 Usage Example (Workflow)

1. Prepare ControlNet frames (all input images).

2. Use Iter Control Group:

   • Iter 1 → input ControlNet → KSampler → output sample_op_1.
   • Iter 2 → input ControlNet + prev_output=sample_op_1 → KSampler → sample_op_2.
   • Iter 3 → input ControlNet + prev_output=sample_op_2 … continue until last iteration.

3. (Optional) Use Debug Number Batch on each sample to confirm frame continuity.

4. Merge all sampler outputs with Merge De-Overlap:

   • (image_1=sample_op_1, image_2=sample_op_2, etc).

5. Send merged sequence to your video writer node.

---

Result: 🎬 A seamless, artifact-free long video.

🔑 Features Recap

• ✅ Designed for WAN v2v long video workflows
• ✅ Handles chunking + overlap seamlessly
• ✅ Prevents skipped or duplicated frames
• ✅ Debug-friendly with numbered overlays
• ✅ Flexible: supports up to 10 chunk merges in one step

📌 SEO Keywords

• ComfyUI
• WAN v2v
• video stitching
• long video generation
• chunk overlap
• ControlNet manager
• de-overlap merge
• KSampler workflow
• seamless video
• AI video tools

📜 License

MIT License – free to use and modify.