Nodes/BizyAirPlus/Vidu Q3 Pro Start End To Video (Channel Edition)
ComfyUI Node

Vidu Q3 Pro Start End To Video (Channel Edition)

Vidu Q3 Pro first-last-frame, the cheap channel edition — bookend your shots

By siliconflow·Created 5 months ago·Updated about a month ago· 20
Vidu Q3 Pro Start End To Video (Channel Edition)
  • first_frame_image
  • last_frame_image
  • api_config
  • video
  • request_id
  • urls
prompt
resolution720P
duration5
stylegeneral
movement_amplitudeauto
audiotrue
skip_errorfalse

Vidu Q3 is the API-only video model that has the community reaching for superlatives about character expressions, and the Pro tier is the quality one. This node is the "Start End To Video" flavor in the Channel (Base) edition - the leaner, typically cheaper route to the same model, with the knobs trimmed to just what a shot needs.

What this node does

BizyTRD_Vidu_Q3_Pro_Base_FLF2V_API animates between a first and last frame. You hand it a start image, an end image, and a prompt, and it invents the motion in between - the tool for match cuts, locked-in shot endings, and storyboard in-betweening. The Channel edition keeps this focused: no seed control, no recommendation-prompt upsell, no off-peak mode. Just the shot.

It's cloud execution like every BizyTRD node: flip the BizyAirPlus button ON, the model runs on BizyAir's servers, and ComfyUI streams back the clip. Nothing runs on your GPU.

The inputs that matter

  • first_frame_image / last_frame_image - the two bookends, both required.
  • prompt - 1–4000 characters describing the motion between them.
  • resolution - 720P by default.
  • duration - a combo, 1–16 seconds.
  • style - general by default, with options for different looks.
  • movement_amplitude - auto by default; dial it down if the model is getting too twitchy between your anchors, up if it's too stiff.
  • audio - on by default. Q3 syncs dialogue and sound effects to the video; flip off for silent.

Shared BizyAir extras: api_config (per-request endpoint/key override) and skip_error (error placeholder instead of workflow abort).

Outputs

  • video - the finished clip; wire into a preview or save node.
  • request_id - cloud task ID for logs and support.
  • urls - hosted URLs for the result.

How to install

cd /path/to/ComfyUI/custom_nodes
git clone https://github.com/siliconflow/BizyAirPlus.git
cd BizyAirPlus
python -m pip install -r requirements.txt

Restart ComfyUI (or install "BizyAirPlus" via ComfyUI Manager). Deps: bizyair-cloudberry and bizytrd; no model files. First run: account at bizyair.ai → API key → enter when the ON button prompts (Settings > BizyAirPlus > API Key; key at ~/.BizyAirPlus/apikey.ini).

Where people get burned

First-last-frame is the most sensitive mode to mismatched anchors: if your start and end frames are wildly different, the model has to invent a bridge and it usually shows. Keep the two frames visually related - same subject, related composition - and let the prompt carry the transition.

The other one is pricing. The Channel (Base) edition exists to be the cheaper route, but remember every queue is a billed cloud task; if you're iterating on the motion, iterate at low resolution and short duration first. And the pack gotcha: BizyAir's malware reputation is leftover from a 2025 ComfyUI Manager auto-install bug - it's a legitimate SiliconFlow pack, and if you remove it, do it with pip uninstall bizyair-cloudberry bizytrd, not by deleting the folder.

CategoryBizyAirModelServices/BizyTRD/Vidu

Inputs (10)

NameTypeDefaultDescription
first_frame_imageIMAGEInput the first frame picture as the start frame of the video generation
last_frame_imageIMAGEEnter the last frame picture, as the end of the video generated frame
promptSTRINGVideo generation prompts, 1-4000 characters
resolutionCOMBO720POutput resolution
durationCOMBO5Video length, in seconds, q3-pro available range 1-16 seconds
styleCOMBOgeneralStyle of video generation
movement_amplitudeCOMBOautoThe range of motion of an object in a picture
audioBOOLEANtrueWhether audio and video synchronization generation is enabled. True: outputs video with dialogue and sound effects; false: outputs silent video
api_configoptBIZYAIR_OPENAPI_CONFIG单次请求覆盖base_url和api_key
skip_erroroptBOOLEANfalse开启后遇到错误不中断工作流,输出对应类型的错误占位符

Outputs (3)

NameTypeDescription
videoVIDEO
request_idSTRING
urlsSTRING