Skip to main content
Load Video

Load Video · screenshot coming soon

​
What this node does

Load Video is a ComfyTV Input stage. It does not run AI models—it registers a video file from input/ (or uploaded into that folder) as a project snapshot of type COMFYTV_VIDEO. The node body previews the selected clip. Downstream stages such as Video Clip, Extract Frame, and Demux read that snapshot. Unlike Load Video from Asset, this targets raw files on disk; the asset loader browses the current project’s library from stage runs and sidebar imports.

​
When to use it

  • You have MP4/WebM/MOV footage and want step-by-step ComfyTV editing inside ComfyUI.
  • You exported a clip from another editor into input/ and select it here.
  • You need source footage before native ComfyTV video tools (most pipelines start from images instead).
  • You bridge native ComfyUI VIDEO tensors into ComfyTV (see Types).

​
How it works (why ComfyTV is designed this way)

  • Stage vs Run: No ▶ Run. Dropdown selection or upload updates output instantly without queuing the full graph.
  • Snapshots: The chosen URL is stored in the project; downstream Runs use it. Re-select here if you change the source file.
  • No workflow: File registration only—no workflows/video/ generator backend.

​
Types (COMFYTV_* vs native ComfyUI)

Conversion:
  • Native → ComfyTV: → ComfyTV Video (Run to snapshot)
  • ComfyTV → native: ← ComfyTV Video
Details: bridges.md

​
UI and parameters

​
video

  • What: Dropdown of video files in input/ plus upload.
  • What to set: Pick an existing file or upload (writes to input/).
  • Effect: Defines the source for the whole downstream chain; use common containers (MP4, etc.).
  • Common mistake: Looking here for Video Stage outputs—use Load Video from Asset instead.

​
project_id / parent_output_id (internal)

  • Hidden; maintained by Project and graph wiring.

​
Outputs

​
Step by step for beginners

  1. Add Node → ComfyTV → Input → Load Video.
  2. Add a Project node (recommended).
  3. Place files in input/ or upload on this node.
  4. Select in the dropdown; confirm preview—no Run needed.
  5. Wire video to e.g. Video Clip and set in/out points.
  6. ▶ Run the downstream stage.
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: Load Video vs Load Video from Asset?
A: This node = raw input/ files. Asset loader = project library after generation/import. Q: Type mismatch downstream?
A: Sockets must expect COMFYTV_VIDEO. Native VIDEO needs → ComfyTV Video Bridge. Q: vs Video Stage (AI generation)?
A: This only loads existing files—no text/image-to-video. Use Generate → Video Stage to create clips. Q: Upload not appearing in dropdown?
A: Refresh the page; confirm the extension is recognized as video.
  • Load Video from Asset — project library picker.
  • Load Audio — audio Input counterpart.
  • Video Clip / Extract Frame — common first downstream steps.
  • → ComfyTV Video (Bridge) — import native tensors.