Video Builder
Plan, price, and generate long multi-shot videos from a single prompt or reference.
Written By Leonid
Last updated About 3 hours ago
What it does
Video Builder turns one idea into a finished multi-shot video from 15 seconds to 3 minutes long. Describe what you want to make, and it writes a plan — scenes, shots, a recurring cast, and a price for every part — before anything is generated.
Stills come first and video second, so you can fix faces and framing cheaply before paying for the expensive part — filming.

Quick start
Describe your video in What do you want to make?
Set the length and aspect ratio, then click Plan the video
Check the plan and price, then click Generate
Review the stills, then click Looks good — film it
When all clips are done, click Assemble the video
Examples
A video made with Video Builder, with the brief and settings used. Copy the brief to start from something close to what you need, then edit it. This is the project shown step by step in the guide below.
Example 1 — A Late Night at Home
Example 1 — A Late Night at Home
Brief: A couple comes home late and shares a passionate moment in the bedroom. Four shots — one for the doorway, one for the first kiss against the wall, and two on the bed. Shot 1: Daniel and Elena, a couple in their late twenties, step through the bedroom doorway in the dark, her hand pulling him in by his shirt, both slightly out of breath and smiling. Shot 2: they stop just inside, he presses her gently back against the wall and they kiss slowly and deeply, his hand at her waist, her fingers in his hair. Shot 3: they sink down onto the bed together, still kissing, her hair spilling across the sheets as he leans over her. Shot 4: a close, intimate shot of the two of them kissing on the bed, unhurried and passionate, hands moving along a shoulder and a jawline, foreheads touching between kisses. Two characters throughout — Elena, a beautiful woman in her late twenties with long dark hair and warm skin, wearing a silk slip dress; and Daniel, a man in his late twenties with short dark hair and light stubble, in an unbuttoned white shirt. Warm low light from a bedside lamp and city glow through the window, soft shadows, cinematic and sensual: slow push-ins and close handheld framing. Intimate romantic mood, no dialogue.
Settings: 40s · 3:4 · WAN 3.0 NSFW · 720p
Step-by-step guide
Every video goes through five stages.
Brief
Step 1. Open Video Builder from the Apps section.
Step 2. Describe your video. Type your idea into What do you want to make? (up to 4,000 characters). The more you specify — the story, the scenes, the number of shots, how the main character looks — the closer the plan will be to what you have in mind.
Step 3. Add a reference (optional). Click Add an image or video:
Images — up to 8. The first image becomes the Style reference for your character; further images are saved as Frames you can use as a shot's opening picture.
Video — exactly 1, and it replaces any attached images. Every shot then copies the movement from your video (motion transfer).
Files can be up to 50 MB, and the shortest side must be at least 512 px.
Step 4. Set the length and aspect ratio. Length goes from 15 to 180 seconds in 5-second steps (default 60s; best results between 30 seconds and 2 minutes). Aspect offers 16:9, 9:16 (default), 1:1, 21:9, 3:4, and 4:3.
Step 5. Click Plan the video. The planner writes the title, scenes, shots, and cast, plus a price quote. Nothing is filmed yet.
Writing a good brief
Tell a story — what happens from beginning to end, not just a single image.
Name the scenes and shots — e.g. "Three scenes: …, …, and …. Six shots."
Describe your character once, in detail — appearance, hair, and clothing. The planner builds a portrait from it and keeps the character consistent across shots.
Set the mood — location, time of day, light, and camera style (tracking shots, slow push-ins, wide establishing shots).
Plan & quote
Step 6. Review the plan. You'll see the video's title, a note on how the brief was split into shots, a description of each recurring character, and every shot grouped by scene.
Step 7. Edit the shots (optional). Every shot is a card you can change — the motion, the opening frame, the model, length, and resolution. Use Add shot after, Start a scene here, Delete shot, and Delete scene to change the structure. Each edit updates the price. See Shot settings below for every field.
Step 8. Check the price and generate. The quote lists every part of the video with its cost. Three optional add-ons sit below it:
Pin faces to the portrait — puts the character's real portrait face into every still instead of relying on the image editor to keep it.
Offer a second camera angle — makes an alternative camera angle for every shot after the first, so you can pick between them before filming.
Generate sound — voices, ambience, and effects made by the video model along with each clip. It's free on the models that support it, and the toggle only appears when every shot in the plan uses one.
When you're happy, click Generate.
Check your balance first. The app can't check your credits before you press Generate. If you run out mid-render, top up and click Resume render — finished shots aren't paid for twice.
The look
Step 9. Review the stills. One still is generated for every shot, plus one portrait per recurring character. Filming is the expensive part, so check every still first and fix anything you don't like:
Step 10. Click Looks good — film it. After this, the plan locks except for prompts, model, and resolution.
Fix it here, not later. Try another look costs 1 credit, while re-filming a clip costs many times more.
Filming
Step 11. Wait for the clips. Clips are generated three at a time; each usually takes around 2–3 minutes. The counter at the top shows how many shots are finished, and every card shows its status:
If a finished shot isn't right, edit its motion prompt and click Regenerate. The price is shown on the button, and the old clip stays until the new one is ready.
Assemble
Step 12. Set the final look. Once every shot has a finished clip, choose:
Film look — None, Film (soft grain, lower contrast), Warm (amber shadows), or Cool (cold, desaturated, more grain).
Title and chapter cards — an opening title, cards before each scene, and a closing card on a background made for your video. Edit the text, pick one of six typefaces, click Rewrite for new wording, or New background for a new background.
Step 13. Click Assemble the video. The clips are joined into one video.
Step 14. Watch and download. When you see Your video is ready, play it and download it from the player's three-dot menu. To change the look or cards later, open Adjust and re-assemble — it reuses your clips, so nothing is filmed again.
All your projects are listed under Your videos at the top of the app. Click one to reopen it.
Shot settings
Every shot in the plan is a card with the settings below. Click a setting to see what it does.
Motion
Motion
What happens in the clip. Describe a movement, not a single moment — e.g. "she turns from the rail and walks toward the door", not "a woman at the rail". Up to 2,000 characters.
Opening frame
Opening frame
The still the clip starts from: framing, clothes, light. Leave it blank to build the still from the motion text.
What has changed about them
What has changed about them
How the character differs from their portrait from this shot on — e.g. coat off, hair wet. It carries forward to every later shot, so write it once.
Video model
Video model
19 models to choose from, set per shot; the line under the picker describes the model you chose. Models marked 18+ require mature-content access on your account. Changing the model can adjust the shot's length and resolution to what that model supports.
Length
Length
5, 6, 8, 10, 12, or 15 seconds; 20, 25, and 30 seconds are available on WAN 3.0 only (WAN 3.0 doesn't support 21:9). Changing a shot's length rebalances the other shots to keep the total close to your target.
Resolution
Resolution
480p, 720p, or 1080p, depending on the model. Motion-transfer shots go up to 720p.
Who appears
Who appears
Which recurring character is in the shot. Choose No one for landscapes and objects, or Everyone when there are several characters.
Opening frame image
Opening frame image
Use one of your uploaded pictures as the shot's first frame, exactly as uploaded. It's free, but Try another look and Fix the face aren't available for that shot.
Style frame
Style frame
Appears only if you uploaded a reference video — picks which of its frames sets this shot's look.
Continue from shot
Continue from shot
The shot opens on the exact last frame of the previous clip, so the two play as one continuous move. It needs no still of its own, so it's cheaper — but its Opening frame text is ignored, so turn it off when the shot needs its own framing or a new location. Not available on the first shot of a scene.
What you can edit, and when. Everything is editable during Plan & quote and The look. Once filming starts, only the model, resolution, and prompts can change — length and character state are locked because they affect shots that are already filmed. Nothing can be edited while a render is running.
Costs
The estimate stays at the top of every screen, split into Images, Videos, and Combining. The Charged so far panel lists what the video has actually cost, line by line.
For example, WAN 2.7 costs 10 / 20 / 30 credits for a 5 / 10 / 15-second clip at 720p, and 15 / 30 / 45 at 1080p.
A plain number is a fixed price; a tilde (~) or a range is an estimate — clip prices are only known exactly once billed. Credits are charged as each step starts, and a failed generation is still charged.
Pro tips
Chain shots for smooth motion — Continue from shot turns two clips into one continuous move and saves the cost of a still.
Write state changes once — if your character takes off a coat in shot 3, note it in What has changed about them on shot 3; it carries into every later shot.
Use No one for establishing shots — landscapes and object close-ups don't need a character.
Start short — try a 30-second version first to test your brief, then build the full-length video.
Troubleshooting
A shot failed because the content was refused — if your video is meant to be adult, switch its shots to an 18+ model (e.g. WAN 2.7 NSFW), which requires mature-content access. If it isn't, reword the prompt — something in it read as explicit. Content involving minors, real people, or non-consent is never allowed.
Shots are marked Held — fix and regenerate the failed shot before them, then resume; everything already filmed is kept.
"This render stopped before finishing" — click Resume render. Only unfinished shots are generated and charged.
You ran out of credits — top up your balance, then click Resume render.
The plan is a few seconds short of your target — the shortest possible shot is 5 seconds, so small gaps can't always be filled. Adjust shot lengths manually if needed.
An 18+ model is refused — these models require mature-content access on your account.
"Output link expired — reopen the app" — video links last 12 hours; reopen Video Builder to refresh them.
Fullscreen doesn't work — the player runs inside the platform, so fullscreen isn't available. Download the video and play it on your device.
A video can't be deleted — it's still rendering; wait until it finishes.
Sending a task ID to support. Every generation Video Builder runs — each still and each clip — is also saved in the tool that produced it. Open that tool and look under Tasks: every generation is listed there with its own task ID. If you need help with a specific shot, include that task ID in your message — it lets our support team look up that exact generation.
Results from AI models can vary and may not always be fully accurate. If you continue to experience issues, contact our support team at [email protected].
FAQ
How much does a video cost?
How much does a video cost?
It depends on the number of shots, the models, resolution, and clip lengths. You get a full quote before anything is generated, and the estimate stays at the top of every screen. Clips make up most of the cost.
How long can my video be?
How long can my video be?
From 15 seconds to 3 minutes. It works best for videos between 30 seconds and 2 minutes.
Can I use my own character?
Can I use my own character?
Yes. Upload a photo as a reference — the first image becomes your character's style reference, and your first portrait is free. The image's shortest side must be at least 512 px.
Can I rename my video?
Can I rename my video?
No. The title is set by the planner. To get a different title, plan the video again from a new brief.
Will I be charged again if I re-assemble?
Will I be charged again if I re-assemble?
Re-assembling only redoes the join, film look, and title cards, so no shot is filmed again. Joining the clips costs 2 credits. Extras are charged separately: title cards cost 2 credits the first time, Rewrite 2 credits, and New background 1 credit. Film look and typeface are free.
Am I charged for a failed shot?
Am I charged for a failed shot?
Yes. Credits are charged when generation starts, so a failed attempt is still billed. Shots that are Held or Interrupted aren't charged until they're generated.
How do I delete a video?
How do I delete a video?
Open Your videos and click Delete next to it. There's no confirmation step, so the video and its plan are removed immediately.
Do the clips have sound?
Do the clips have sound?
It depends on the model. Models that support sound generate voices, ambience, and effects along with the clip at no extra cost — turn on Generate sound in the plan. On other models the video is silent; you can create a soundtrack with Zen Audio and add it in any video editor. Shot previews in the app always play muted.
Why do the voices change between shots?
Why do the voices change between shots?
Each shot's sound is generated on its own, without hearing the shots around it, so a character can sound different from one shot to the next and the background can change at every cut.
What's next
Video Analyzer — break down a reference video before building your own
Image to Video — animate a single image
Pricing & Credits — see credit costs across all tools