Nodes/BizyAirPlus/Vidu Q3 Pro Image To Video (Official Edition)
ComfyUI Node

Vidu Q3 Pro Image To Video (Official Edition)

Vidu Q3 Pro image-to-video, Official edition — seed-locked takes and the optional end frame

By siliconflow·Created 5 months ago·Updated about a month ago· 20
Vidu Q3 Pro Image To Video (Official Edition)
  • image
  • last_frame_image
  • api_config
  • video
  • request_id
  • urls
prompt
resolution720P
aspect_ratio16:9
duration5
audiotrue
seed0
is_recfalse
skip_errorfalse

If you're running a serious Vidu Q3 Pro workflow, this is the one you'll actually live in. The Official I2V node keeps everything that made the Channel version pleasant - one image, one prompt - and adds the two things heavy users need: a seed, and an optional end frame that turns the shot into first-last-frame on demand.

What this node does

BizyTRD_Vidu_Q3_Pro_Official_I2V_API animates an image first frame forward according to your prompt. It's the workhorse of the Q3 Pro Official family, and the tooltip notes your input image's aspect ratio should be between 1:4 and 4:1 - outside that band the model gets awkward.

Cloud execution as always: flip BizyAirPlus ON, and Q3 Pro runs on BizyAir's servers; ComfyUI uploads your image and streams back the finished clip.

The inputs that matter

  • image - the first frame. Feed it a good still; Q3 is strong enough that your input quality is the ceiling.
  • prompt - up to 5000 chars of motion.
  • resolution (720P default), aspect_ratio (16:9 default, inherits from image when unset), duration (1–16s).
  • audio - on by default: synchronized dialogue and SFX.
  • seed - 0 default, -1 = random. The reason to pick this node over the Channel edition: lock a seed and you can iterate on prompts while keeping a take reproducible.
  • last_frame_image - the subtle superpower. Optional here; provide one and this I2V node silently becomes a first-last-frame generator. No node swap needed.
  • is_rec - recommendation prompt at an extra 10 credits per task. Much saner than the 320-coin version on the FLF node, but still a paid upsell - leave off unless you want auto-prompting.

Plus the shared api_config and skip_error options.

Outputs

  • video - the finished clip, ready for preview/save.
  • request_id - cloud task ID.
  • urls - hosted result URLs.

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 (or Settings > BizyAirPlus; key at ~/.BizyAirPlus/apikey.ini).

Where people get burned

The seed semantics catch people: 0 is a real fixed seed, so "I changed nothing and got the same result" is expected behavior, and -1 is what you want for variety. The aspect-ratio band on the input image (1:4 to 4:1) is the other one - an ultrawide or tall input outside the band tends to produce weird crops.

And the standing pack note: BizyAir's malware reputation is leftover from a 2025 ComfyUI Manager auto-install bug - legitimate SiliconFlow pack, and if you ever remove it, do it with pip uninstall bizyair-cloudberry bizytrd rather than deleting the folder alone.

CategoryBizyAirModelServices/BizyTRD/Vidu

Inputs (11)

NameTypeDefaultDescription
imageIMAGEInput the first frame picture as the start frame of video generation. The aspect ratio should be between 1:4 and 4:1
promptSTRINGVideo generation prompts, maximum 5000 characters. Manually entered is will be ignored if is is enabled
resolutionCOMBO720POutput resolution, viduq3-pro defaults to 720P
aspect_ratioCOMBO16:9Output video aspect ratio, not set by the input picture decision
durationINT51–16Video length, in seconds, viduq3-pro available range 1-16 seconds
audioBOOLEANtrueWhether audio and video synchronization generation is enabled. True: outputs video with dialogue and sound effects; false: outputs silent video. Q3 model support only
last_frame_imageoptIMAGEOptional end frame picture. After providing, it will be used as the end frame of the video, realizing the first and last frame video generation
seedoptINT0-1–2147483647Random seed, -1 means automatically generated
is_recoptBOOLEANfalseWhether to use a recommendation prompt. When enabled, the system automatically generates recommended prompts, which cost an additional 10 credits per task
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