Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.heyoz.com/llms.txt

Use this file to discover all available pages before exploring further.

The available models have been updated since this format launched, so the model names and token costs in this guide reflect the current lineup shown in the interface.

How to Get There

Click “Content Studio” in the left menu. Scroll down on the “What will you create?” screen and select “Video Generator.”

Step 1: Upload a Reference Image (Optional)

The first field on the Custom Video Creator screen asks you to upload an image to reference in your prompt. This step is optional. If you include an image, it becomes the first frame of the generated video. The model uses it as a visual anchor, so the video starts exactly from that scene and evolves based on your prompt. To add one, click “Select Image” to open your Assets Library. Your brand’s product images are organized by product folder. Select the one you want, or drag and drop a new file to upload it. Once selected, you can crop it or edit it with AI before confirming. Click “Use Selected” when ready. Skip this step entirely if you’d rather build the scene from the prompt alone.

Step 2: Write Your Video Prompt

In the second field, describe the video you want to generate. This is where most of the creative direction happens, so be specific. Good prompts include what’s happening in the scene, how the camera moves, what the environment looks like, and any motion or action taking place. A few examples of prompts that work well:
  • For image-referenced videos: “The product begins rotating slowly as soft cinematic light moves across it.” or “The person in the image smiles and leans forward slightly as the camera zooms in.”
  • For text-only generation: “A woman walking through a sunflower field during golden hour, handheld cinematic style.” or “A futuristic drone flying above a neon-lit city at night.”
Keep prompts focused: 10 to 30 words tends to produce cleaner results than long, overloaded descriptions. Add specific visual details (soft warm lighting, marble surface, neon reflections) and camera directions (orbit around, pan left, slow zoom). Avoid vague descriptors like “good,” “cool,” or “nice” as they don’t give the model anything to work with. The Documentation panel on the right has expandable sections covering how to upload an image, how to write a video prompt, and how to select a model. Worth a quick read if you’re new to the format. Click “Continue” when your prompt is ready.

Step 3: Select a Video Generation Model

The next screen shows the available models with their token costs and a short description of each. The current lineup is:
  • Veo 3.1 (64 tokens): Highest realism, cinematic shots, smoothest motion. Best for premium ad content where quality is the priority.
  • Seedance 1.5 Pro (16 tokens): Video and audio generation with lip-sync support. Fast and low-cost, good for quick concept tests or UGC-style content.
  • Wan 2.6 (20 tokens): Artistic video styles. Works well for creative, stylized, or illustrative content.
  • Kling 3 (25 tokens): Cinematic video with audio support. Best for movement-heavy scenes and content where audio matters.
Select one model per generation. You can run multiple generations with different models to compare results. Below the model selection, you’ll see a “Configure Models” section with two settings: Resolution (currently 720p) and Aspect Ratio (Portrait, Landscape, or Square). Set these based on where you’re planning to use the video. The token cost shown on the “Create Videos” button updates based on which model you select. When ready, click “Create Videos.”

Step 4: Review Your Generated Video

HeyOz redirects you to My Creations. You’ll see a “Video generation started successfully” notification at the bottom of the screen. The new video appears at the top of the list with a “Processing” status and an estimated time remaining. Once complete, click the video to open the preview panel. The video plays directly in the panel so you can review it before taking any action.

Step 5: Edit, Post, or Download

From the preview panel you can:
  • Open Video Editor to trim the video, add text overlays, insert subtitles, add more footage, update music or sound effects, and adjust transitions before publishing.
  • Redo Creation to generate a fresh version using the same prompt and settings.
  • Post Now to publish directly to your connected social accounts.
  • Download to save the file to your device.
  • Delete to remove it from your creations.

Tips for Better Results

  • Model choice matters more than prompt length. A short, specific prompt on Veo 3.1 will generally outperform a long, detailed prompt on a faster model if realism is what you need. Match the model to the use case rather than defaulting to the cheapest option every time.
  • If you’re using a reference image, make sure it’s clean and well-lit. The model follows the visual qualities of the first frame closely, so a low-quality input tends to produce a low-quality output.
  • Camera direction in your prompt makes a noticeable difference. “Slow rotation” or “camera pulls back” gives the model something concrete to work with. Without it, the motion can feel arbitrary.
  • Run the same prompt on two different models before committing. The token cost is low enough on Seedance 1.5 Pro and Wan 2.6 that it’s worth comparing before generating a final version on Veo 3.1.
Generated videos are saved automatically to My Creations and can be edited or reposted anytime.
Image4
Image5
Image3
Image2
Image6
Image1
Image8
Image7