🔥AddTextWatermark

AddTextWatermark is the pack's video titling node - it renders a text string onto the video at a chosen position, size, and color. Think lower-thirds, "unedited footage" stamps, or just burning a channel name into the corner. The standout feature here is that it handles Chinese fonts without complaint: the README explicitly shows dropping a .ttf like Alibaba-PuHuiTi-Heavy.ttf into the pack's fonts folder, and the node will happily render CJK text. That's a surprisingly rare thing to find working out of the box in this ecosystem, and it's clearly one of the author's own use cases (the pack's docs are Chinese-first).

How it works

It's FFmpeg's drawtext filter under the hood. The node registers the pack's fonts directory as a ComfyUI search path, so your font choices appear as a dropdown - default, or any .ttf you drop into custom_nodes/ComfyUI-FFmpeg/fonts/. The text is drawn at position_x/position_y (top-left origin) at the given font_size and font_color, then the video is re-encoded. Output is timestamp-named so you never overwrite a previous pass.

The inputs that matter

• video_path / output_path - the usual pair: source video, existing output directory.
• text - what gets drawn. Default Watermark.
• font_file - the dropdown of available .ttf files. This is the one that trips people up: if the list is empty, you haven't put any font in the fonts folder. It's not a bug - the pack ships the folder empty (it contains only a put_fontfile_here placeholder file).
• font_size - pixels, default 15, up to 1000.
• font_color - #FFFFFF style hex or a named color (white, red, …). Default white.
• position_x / position_y - pixel coordinates, default 10, 10.

Outputs

Two strings: video_path and output_path, both pointing at the rendered file. Slightly redundant, but harmless.

Install

Pack-wide standard: ComfyUI Manager → search "ComfyUI-FFmpeg", or manually:

cd ComfyUI/custom_nodes
git clone https://github.com/MoonHugo/ComfyUI-FFmpeg
cd ComfyUI-FFmpeg
pip install -r requirements.txt

Then restart ComfyUI. Remember the real prerequisite is FFmpeg on PATH - the requirements.txt is a single-line affair.

Gotchas

• The empty fonts folder is the number one failure here. Drop a .ttf in, refresh, and the dropdown populates. Windows users can grab a system font from C:\Windows\Fonts.
• No opacity or outline options. A white #FFFFFF label on a white sky is invisible. Pick your color with the video in mind.
• If your text is a single long word that overflows the frame edge, it just gets clipped - there's no auto-wrap or size-to-fit. Keep captions short.