Video Generation
Generate brand-aware short-form videos with AI, directly inside Cuppa. Choose from multiple video models, control aspect ratio, duration, and resolution, then share to social or download for use anywhere.
What You Can Do
- Generate videos from text prompts: Describe a scene and let AI create it
- Use the Video Script Generator: Turn a simple brief into a detailed, cinematic prompt
- Use images as start frames: Pick any image from your Media Library as the first frame
- Brand-aware generation: Cuppa automatically weaves in your brand colors, style, and industry context
- Multiple AI models: Seedance 2.0, Runway Gen-4.5, Grok Imagine Video, Kling v3, Veo 3.1 Fast, P-Video, Higgsfield DoP, Wan 2.7 (WaveSpeed)
- Model-specific controls: Duration, FPS, aspect ratio, resolution, negative prompts, seeds, and more
- Share to social: Publish directly to LinkedIn, X, and other connected platforms
- Download: Save videos locally in full quality
Getting There
Navigate to Creative Studio > Video Generation in the left sidebar. You’ll see the generation form at the top and a gallery of your generated videos below.
Choosing a Model
Click the model selector at the bottom-left of the prompt area. Each model has different strengths:
| Model | Best For | Duration | Key Features |
|---|---|---|---|
| Seedance 2.0 | Multimodal, native audio, video editing | 5-15s | Audio sync, image/audio refs, video extension |
| Seedance 2.0 Fast | Same as Seedance 2.0, faster generation | 5-15s | Same capabilities, optimized for speed |
| Runway Gen-4.5 | Benchmark quality, cinematic | 5-10s | Seed, 720p/1080p |
| Grok Imagine Video | Audio generation, image-to-video | 5-15s | Audio sync, start image, editing |
| Kling v3 | Multi-shot, detailed scenes | 3-15s | Audio, negative prompt, start/end image |
| Veo 3.1 Fast | Quick generation, Google quality | 4-8s | Audio, negative prompt, seed |
| P-Video | Creative/artistic, audio-driven | 1-10s | Audio input, FPS control, draft mode |
| Higgsfield DoP | Cinematic camera moves on a still image | 5s | Image-to-video, start/end frame, 720p |
| Wan 2.7 (WaveSpeed) | Text or image to video | 5-15s | Text-to-video default; pass start image for I2V, 720p/1080p |
Generation Settings
Settings change based on which model you select. Here’s what’s available:
Universal Settings
- Duration: How long the video will be (model-specific steps)
- Aspect Ratio: 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 1:1 (varies by model)
- Resolution: 720p or 1080p (Sora 2 Pro only supports 1080p)
Model-Specific Settings
| Setting | Available On | Description |
|---|---|---|
| Start Image | Seedance (both), Grok, Kling, Veo, P-Video, Higgsfield DoP, Wan 2.7 | Use a Media Library image as the first frame (REQUIRED for Higgsfield DoP; switches Wan 2.7 to image-to-video) |
| End Image | Seedance (both), Kling v3, Veo 3.1, Higgsfield DoP, Wan 2.7 | Create transitions between two images |
| Audio Input | Seedance (both), P-Video | Reference audio for rhythm/sound matching |
| Negative Prompt | Kling, Veo | Describe what you don’t want in the video |
| FPS | P-Video | 24fps or 48fps toggle |
| Draft Mode | P-Video | 4x faster preview quality |
| Prompt Upsampling | P-Video | Auto-enhance your prompt |
| Seed | Seedance (both), Runway, Grok, Veo, P-Video | Reproducible results |
| Audio Generation | Seedance (both), Grok, Kling, Veo | Native audio sync |
Video Script Generator
Don’t want to write a detailed prompt? Use the Script Generator to turn a simple brief into a professional, cinematic prompt.
How to Use
- Click Script Generator in the prompt area
- Choose a category (Product Showcase, Brand Story, Social Short, Tutorial, etc.)
- Enter a brief description of what you want
- Cuppa generates a detailed prompt with camera movements, lighting, pacing, and brand context
- Review and edit, then generate
The Script Generator uses your Brand DNA (colors, industry, voice) to tailor every prompt.
Brand-Aware Prompts
When you have Brand DNA configured for your selected brand, Cuppa automatically enriches your prompt with:
- Brand colors as natural color references (e.g., “fresh green accent lighting”)
- Industry-specific motion language (e.g., clean dolly moves for SaaS, handheld energy for food)
- Visual quality standards (cinematic lighting, texture fidelity, film-like tonality)
This happens transparently. Your prompt stays clean while the generated video reflects your brand.
Generation Time
Video generation takes longer than images. Expect:
| Model | Typical Time |
|---|---|
| Veo 3.1 Fast | 1-4 minutes |
| Grok Imagine Video | 1-4 minutes |
| Higgsfield DoP | 1-3 minutes |
| Runway Gen-4.5 | 2-6 minutes |
| P-Video | 2-6 minutes |
| Kling v3 | 2-8 minutes |
| Seedance 2.0 | 2-8 minutes |
| Seedance 2.0 Fast | 1-5 minutes |
A loading card appears in the gallery immediately after you hit Generate. You can navigate away and come back; generation continues in the background.
Video Gallery
Generated videos appear in a gallery below the form, sorted by most recent.
Video Card Actions
Hover over any completed video to see quick actions:
| Action | Description |
|---|---|
| Play | Open fullscreen lightbox to watch the video |
| Download | Save to your computer |
| Share to Social | Publish to connected social platforms |
| Edit & Retry | Prefill the form with this video’s settings for a new generation |
| Show Full Prompt | Expand to see the complete prompt used |
Statuses
| Status | What It Means |
|---|---|
| Creating your video… | Generation in progress (green loading card) |
| Complete | Video is ready to play, download, or share |
| Failed | Generation failed. Click Edit & Retry to try again |
Media Library Integration
All generated videos also appear in your Media Library alongside images. You can filter by All, Images, or Videos. From the Media Library, you can:
- Play videos in a detail modal
- Download
- Share to social
- View metadata (model, prompt, duration, resolution)
API Keys Required
Video generation uses your own API keys (BYOK model):
| Model | API Key Needed |
|---|---|
| Seedance 2.0 / 2.0 Fast, Runway Gen-4.5, Grok, Kling, Veo, P-Video | Replicate API key |
| Higgsfield DoP | WaveSpeed API key |
Configure API keys in Team Settings > API Keys. You can hold multiple keys for both providers; Cuppa picks the right one based on the model you choose.
Tip: Higgsfield DoP is image-to-video - it always needs a start image. Pick one from your Media Library or paste a public URL. The DoP model excels at director-style camera moves (orbit, dolly, parallax push, vertigo zoom) on a still you already love.
Plan Availability
| Feature | Minimum Plan |
|---|---|
| Video Generation | Studio |
| Video Script Generator | Studio |
| Share to Social | Studio |
| Brand-Aware Prompts | Studio (with Brand DNA configured) |
Troubleshooting
Video stuck on “Creating your video…”
- Generation can take up to 8 minutes for some models
- If stuck beyond 15 minutes, the job will auto-fail
- Check that your API key is valid and has credits/quota
”Failed” generation
- Click Edit & Retry to try again with the same settings
- Try a different model or simplify your prompt
- Some prompts may be blocked by the model’s content moderation. Simpler, more descriptive prompts tend to pass
Video quality not what you expected
- Use the Video Script Generator for more detailed, cinematic prompts
- Configure Brand DNA for automatic style enrichment
- Try Runway Gen-4.5 at 1080p or Seedance 2.0 at 720p for highest quality
- Adjust duration: longer videos give the AI more room for scene development
Hex codes or technical text appearing in video
- This is handled automatically by Cuppa’s prompt compression
- If you see artifacts, try regenerating. The compression step converts all technical references to natural language
Related Docs
- Brand Asset Library: Upload and generate images
- Mood Board: Visual inspiration board
- Social Connections: Connect platforms for sharing
- Brand Kit: Brand colors and visual identity