Nodes/ComfyUI Beeble AI/Beeble Upload
ComfyUI Node

Beeble Upload

Pre-stage a source asset on Beeble storage and get back its beeble_uri. Use when one source feeds many SwitchX runs (prompt sweeps, alpha-mode comparisons) so the same bytes don't get re-uploaded for every generation. For one-shot SwitchX runs, you don't need this -- the SwitchX node uploads internally. Output is a STRING beeble:// URI you feed into SwitchX's source_uri / reference_image_uri / alpha_uri.

By am-pipeline-prod·Created 3 months ago·Updated 3 months ago· 3
Beeble Upload
  • source
  • video
  • beeble_uri
media_typeimage
filenameinput.png
fps24.0
CategoryBeeble AI

Inputs (5)

NameTypeDefaultDescription
sourceIMAGEThe source asset to upload. Single still for media_type='image' (PNG); multi-frame batch for media_type='video' (MP4). Same data shape as the Beeble SwitchX 'source' / 'alpha' / 'reference_image' inputs.
media_typeCOMBOimageHow to encode and upload the IMAGE input. 'image': single still, written as PNG. 'video': frame batch, encoded as MP4 (libx264, yuv420p).
filenameSTRINGinput.pngFilename hint sent to the API. Extension is auto-corrected to .png or .mp4 to match media_type, so this is mostly cosmetic for the Beeble dashboard.
fpsoptFLOAT24.01–240Frames per second (only used when media_type='video').
videooptVIDEOOptional VIDEO input. When wired, this node uploads the underlying file directly (for `VideoFromFile` sources) or PyAV-encodes to a temp MP4 (for `VideoFromComponents` / lazy AM transform chains) — no IMAGE-batch materialisation, no re-encode from uint8. media_type is forced to 'video' and `source` is ignored. Wire this when upstream is Topaz Video Enhance, AM Read Video → VIDEO socket, or a lazy AM transform chain.

Outputs (1)

NameTypeDescription
beeble_uriSTRING