Skip to main content
Subtitles

Subtitles · screenshot coming soon

​
What this node does

Subtitles takes SRT/WebVTT cue text and burns it, cue by cue with the right timing, onto a COMFYTV_VIDEO. You paste subtitle text (or wire a COMFYTV_TEXT input), style the font, then press ▶ Run; it renders with an ffmpeg pass on the server and returns a new COMFYTV_VIDEO. You can supply the cues two ways: paste into the subs text box, or connect a subs_text (COMFYTV_TEXT) input — for example the output of Subtitles · Speech-to-Text. If both are present, the wired subs_text wins. The node needs a video and non-empty cue text, or Run errors.

​
When to use it

  • Add open captions (baked-in subtitles) to a clip for social platforms.
  • Turn a transcript from Subtitles · Speech-to-Text into on-screen captions in one wire.
  • Localize a clip by pasting a translated SRT.

​
Parameters

​
subs

Multiline SRT or WebVTT cue text, pasted directly. Used if no subs_text input is wired.

​
font

Dropdown of server fonts (from list_fonts(), falls back to default). Default is the first installed font.

​
size

Caption font size in pixels. Range 8–200, default 36.

​
color

Caption fill color, hex string. Default #FFFFFF.

​
stroke

Outline thickness in pixels, 0–20, default 2. A small stroke keeps captions readable.

​
anchor

Caption placement: bottom or top. Default bottom.

​
Inputs and outputs

​
Tips

  • Wire Subtitles · Speech-to-Text → subs_text to go from audio to on-screen captions without copy-paste.
  • The wired subs_text takes priority over the pasted subs box; clear the box if you meant to use the wire.
  • Cue timing comes from the SRT/VTT itself — this node does not re-time cues, it renders them as given.

​
Types and bridges

COMFYTV_VIDEO and COMFYTV_TEXT are ComfyTV project snapshots, not native ComfyUI types. To interoperate with native ComfyUI nodes, insert a Bridge (ComfyTV/Bridge). See bridges.md.
  • Subtitles · Speech-to-Text — generate the SRT/VTT cues from speech to feed subs_text.
  • Title — a single styled title card instead of many timed cues.