Skip to main content
Subtitle Erase · Smart

Subtitle Erase · Smart · screenshot coming soon

​
Intended purpose

Subtitle Erase (Smart) targets hard-coded subtitles burned into pixels: detect + inpaint each frame → clean COMFYTV_VIDEO. No manual box — good when subtitle placement is consistent (anime clips, interviews).
⚠️ Current status: placeholder only. Run → StageNotImplemented. Region variant + box UI also pending. See roadmap.md.

​
When to use it (once shipped)

  • Remove subs before re-edit or I2V
  • Batch stock footage with fixed lower-third text
  • Fallback to Subtitle Erase (Region) when Smart fails

​
Planned design

  • Stage + snapshot; COMFYTV_VIDEO in/out.
  • OCR/detection + video inpaint workflow (TBD).
  • Workaround: Video Crop bottom band; or external tool → → ComfyTV Video.

​
Types

​
Parameters

​
video

Source clip (required when implemented). Wire only for layout today.

​
Outputs (when implemented)

​
Step by step (today)

  1. Node not available.
  2. Bottom-only subs: try Video Crop.
  3. Watch roadmap.
This page covers one node only. For end-to-end workflows, multi-stage pipelines, type conversion, and design rationale, see the ComfyTV user guides on GitHub:

​
Repository and workflows

​
FAQ

Q: vs Region?
A: Smart auto-detects; Region manual rectangle (per-frame UI pending). Q: Soft subs in mkv?
A: Container tracks — strip with Demux / ffmpeg, not this node. Q: Run error expected?
A: Yes — not implemented.
  • Subtitle Erase (Region) (pending)
  • Video Crop (✅)
  • → ComfyTV Video