Skip to content

Product Ad in One Click

Template
Product Ad in One Click, in the ComfyUI template browser under BytePlus
Workflow file
Product Ad in One Click.json
BytePlus nodes
BytePlus LLM, Seedream 4.5 & 5.0, Seedance 2.5 First-Last-Frame to Video, vCube Video Enhance
Keys needed
ModelArk API key, AI MediaKit API key
Also uses
Extract Text, MarkdownNote, Preview as Text, Save Video, Text (Multiline) (built into ComfyUI)

A short 16:9 product ad in 1080p, made from nothing but a sentence. It is for makers who have a product idea but no photo yet. The idea: let a language model act as the creative director, write one image prompt and one motion prompt, and hand them to three specialists in turn: Seedream paints the hero shot, Seedance brings it to life and vCube upscales and cleans the final clip.

The Product Ad in One Click template in five groups: brief, creative director, hero shot, animate, enhance and save
Five groups from the brief to the saved 1080p ad.
  1. A Product brief (one line) text node holds the idea. The template uses a new cold-brew coffee in a matte-black can for busy students.
  2. BytePlus LLM (Seed 2.1 Turbo, titled “Creative director”) answers in the format IMAGE: and MOTION:. Two Extract Text nodes split the reply, and two Preview as Text nodes show both prompts.
  3. Seedream 4.5 & 5.0 (Seedream 5.0 Pro, 16:9 at 2048x1152) paints the hero shot from the image prompt.
  4. Seedance 2.5 First-Last-Frame to Video with the model set to Seedance 2.0 animates the image as its first frame: 720p, 5 seconds, audio on.
  5. An Extract Text node pulls the video link out of the Seedance response.
  6. vCube Video Enhance takes that link as video_url and upscales to 1080p. Save Video keeps the result in BytePlus/Video/ProductAd.
  1. Add your ModelArk API key and your AI MediaKit API key in Settings → BytePlus if you have not yet (API keys). The template needs both.
  2. Open the template browser and choose Product Ad in One Click under BytePlus.
  3. Replace the text in Product brief (one line) with your own product and audience.
  4. Press Run. The two prompts appear first, then the ad is saved.
  • Product brief: mention the product, the container or look, and who it is for.
  • model in the hero shot: switch to another Seedream model, for example seedream 5.0 flash, while you tune the brief.
  • duration and generate_audio in Animate: the template makes 5 seconds with audio on.
  • resolution in vCube: 1080p here, with tool_version standard, scene aigc and enhance_style hd.
  • System prompt of the LLM: edit it to change the camera move or the look, for example a handheld style instead of a slow push-in.
Three frames of the saved ad, a matte-black can with condensation against a coffee-bean bokeh background, and the two prompts the LLM wrote
Example result: 1920x1080, about 5 seconds.
The saved ad from this example.

The LLM, Seedream and Seedance nodes bill on your ModelArk account, and vCube bills on your AI MediaKit account. The seed controls are set to randomize, so every run makes new prompts, a new image and a new clip and bills again. Set a control to fixed and ComfyUI reuses the cached result of an unchanged node without billing. vCube reads the Seedance clip from its link, so no Comfy.org login is needed.