This walkthrough takes you from a blank composer to a finished video.
1. Open the composer
Go to the chat and describe the video you want. A single sentence works — “a 30-second ad for a cold-brew coffee brand, energetic, for Instagram” — but the more intent you give, the closer the first cut lands. See Writing great prompts.
Around the composer you’ll find:
- Aspect ratio — 16:9 (widescreen), 9:16 (vertical), 1:1 (square), or 4:5. The visible selection is respected exactly; pick the canvas where the video will actually run.
- Enhance prompt (✨, on by default) — expands a short brief into a fuller creative brief before production, preserving every specific you wrote. Turn it off if you want your text used exactly as-is.
- Video style — optionally pick a format from the style browser (UGC ad, cinematic trailer, explainer, and many more). An explicit pick locks that style’s look. See Video Styles.
- Attachments — add product photos, a person’s photo, a music track, documents, or reference videos. See Uploads & assets.
- Model options — advanced: choose the image/video model and how many image candidates are generated per scene.
2. Say how long it should be
Put the duration in the prompt: “a 30-second ad”, “about 90 seconds”. Notes:
- If you don’t specify, Wavemaker picks a length that fits the content and channel (short-form social defaults to ≤45s).
- Broadcast lengths are frame-exact. Ask for a “:30 TV spot” and the master will be exactly 30.00 seconds — padding or trimming is handled intelligently so the closing card and dialogue are never crushed.
- If you upload a music track and give no duration, the song’s length drives the video.
3. Review the storyboard
Early in the run Wavemaker posts a draft storyboard card: every planned scene with a sketch frame, description, and timing. In normal mode production continues while you look; in step-by-step mode the run pauses here so you can edit scene cards (reorder, reword, remove) and approve before any expensive generation starts.
4. Watch it build
The activity feed shows every step in real time: research, script, storyboard, scene images, quality reviews, clips, narration, music, and assembly. Generation typically takes several minutes for a multi-scene video — clips render in parallel and each passes automated review before it’s accepted.
5. Refine
When the video appears, watch it in the player, then just say what to change:
“Scene 4 feels slow — tighten it.” “Use a warmer voice for the narrator.” “Change the end card to say ‘Order now at coldbrew.co’.”
Each refinement produces a new version — you can always go back. For scene-surgical control, open the Canvas. More in Editing & refinement.
6. Export and share
Use Export for an MP4 (SD/HD/4K tiers), or share the watch page directly. See Exporting & sharing.