Describe a shot, not a story
A clip is four to fifteen seconds long. One clear action with one clear camera behaviour is what fits. Multi-beat narratives are built from several generations joined afterwards, not requested in a single paragraph.
Documentation
FrameAI Guide — AI Video Prompts, Settings & Best Practices
Master FrameAI with this complete guide: prompt structure, reference images, camera language, credit optimization, and the settings that turn a first render into production-ready AI video.
- Generation time varies by model, resolution, duration, and provider load.
- Picture and synchronized audio are generated together in a single pass.
- Commercial use on paid plans, subject to the Terms of Service.
The FrameAI guide — everything between your first render and a usable one
The gap between a first attempt at AI video and something you would actually ship is not talent, and it is not the model. It is a handful of specific habits: describing a shot instead of a story, naming the camera move, using references to hold things constant, and drafting cheaply before committing at full resolution. This guide covers each of them.
Everything here is written against Seedance 2.0 as it actually runs on FrameAI — 4-to-15-second generations at 24 fps, up to 4K, seven aspect ratios, up to nine reference images and three reference clips per request, native audio, and credit costs quoted before you generate. Where a limit is mentioned it is the real limit, enforced by the generator rather than rounded for the marketing copy.
- Prompt structure that consistently works
- When to use each tool, and why
- How to spend a third of the credits
The three habits that change your output most
None of these is advanced. They are simply the things experienced users do automatically and beginners skip.
Name the camera
Unspecified camera work becomes a slow drift. Stating the move and its speed — "slow push-in", "fast whip pan" — is the single highest-value addition to a prompt that is nearly working.
Draft cheap, commit once
Explore at 480p on Seedance 2.0 Mini, then re-run the winner on the flagship at full resolution. Same finished clip, a fraction of the credits, and far more attempts along the way.
Prompt anatomy, and how to debug one that is not working
A reliable prompt has five parts, and you can check them off: subject, action, camera, light, mood. "Slow push-in on a chef plating scallops, tungsten key from the left, shallow depth of field, steam rising" contains all five in one sentence. What it deliberately contains none of is praise — "cinematic", "stunning", "8K", "masterpiece" describe how you feel about the result rather than what should be in it, and the model cannot act on any of them. Replacing every such word with a concrete noun, a lens, or a lighting term is usually the fastest single improvement available.
When a result is close but wrong, resist the urge to rewrite. Change one variable and regenerate, so you learn which word was doing the work. If the composition is wrong, adjust the camera and shot size; if the feeling is wrong, adjust the light; if the subject is wrong, add a reference image rather than more adjectives; if the motion is wrong, adjust the speed qualifier before anything else. And if the model keeps ignoring part of the prompt, it is usually a sign that the prompt is asking for more than fifteen seconds can hold — cut it down and generate the rest as a second shot.
The five-part checklist
Subject, action, camera, light, mood. If a prompt is underperforming, one of the five is usually missing entirely.
Debug by single variable
Composition wrong, change the camera. Feeling wrong, change the light. Subject wrong, add a reference. Motion wrong, change the speed.
Model specifications
What Seedance 2.0 actually delivers on FrameAI
Published limits, not marketing rounding. Every number below is enforced by the generator before your credits are reserved, so a job that would exceed a limit is rejected instead of failing halfway.
- Up to 4K
- Output resolutions — 480p / 720p / 1080p / 4K
- 4–15 sec
- Clip length per generation, at 24 fps
- Native audio
- Sound effects and ambience generated in sync
- 9 + 3
- Reference images plus reference videos, one request
- 7 ratios
- Supported aspect ratios — Auto, 16:9, 4:3, 1:1, 3:4, 9:16, 21:9
- Commercial use on paid plans
- Subject to the Terms of Service
Why creators run Seedance 2.0 on FrameAI
The model is ByteDance’s. What FrameAI adds is the part that decides whether it is usable at work: honest limits, honest billing, and output you are allowed to ship.
Every input in a single pass
Text, up to nine reference images and up to three reference video clips are fused in one generation, not stitched afterwards. There is no separate text-to-video mode to switch into — you add whatever material you have and the model reconciles it.
Sound generated with the picture
Seedance 2.0 writes footsteps, room tone, impacts and score alongside the frames, locked to the action. Most models hand you a silent clip and leave the sound design to you; here the export is already finished.
Seven ratios, up to 4K
One prompt covers a 9:16 vertical cut for Reels, a 16:9 master for YouTube and a 21:9 anamorphic frame for a title sequence. Resolution runs from 480p up to 4K, so the same take can serve a cinema-shaped frame or a phone.
Credits you can predict
The exact credit cost is calculated from your duration, resolution and ratio and shown before you press generate — never after. If a generation fails on the provider side, the reserved credits are returned automatically.
Built for iteration, not for queuing
Seedance 2.0 Fast and Seedance 2.0 Mini exist for the twenty drafts nobody sees. Rough a shot at 480p for a fraction of the credits, then re-run the take that worked at full resolution on the flagship model.
Yours to sell
Paid-plan output may be used commercially under the Terms of Service. Free-credit output is for evaluation only.
A workflow that holds up on real projects
Three phases, in the order that wastes the fewest credits.
- 1
Plan the shots before generating
Write a short shot list — three to five beats, each with a shot size and an action. Knowing what you need is what keeps the credit spend proportional to the result.
- 2
Draft at low resolution
Run each shot on Seedance 2.0 Mini at 480p, changing one variable per attempt. Lock the prompt and the reference set only when a draft reads right.
- 3
Commit, then assemble
Re-run the locked prompts on Seedance 2.0 at 1080p or 4K, then join the shots — in an editor, or with Video Extend and Video Transition.
Which tool for which job
They are all the same model. What differs is what you hand it and what you are trying to hold constant.
Nothing exists yet
Use Text to Video to explore, and Motion Control to make the camera deliberate. This is the right starting point when you are still finding the idea.
Something must stay exact
Use Image to Video for a single subject, or Reference to Video when a character, product and art direction all have to hold at once across several shots.
Footage already exists
Use Video to Video to restyle it, Video Extend to continue it, and Video Transition or First & Last Frames to control how shots begin and end.
Getting started — frequently asked questions
What is the single biggest mistake beginners make?
Writing a story instead of a shot. A four-to-fifteen-second clip can hold one clear action with one clear camera behaviour. Multi-beat sequences are built from several generations joined afterwards, not requested in one paragraph.
How do I stop wasting credits?
Draft at 480p on Seedance 2.0 Mini and change one variable per attempt, so each generation teaches you something. Only re-run on the flagship at full resolution once the prompt and references are locked.
Why does the model ignore part of my prompt?
Usually because the prompt asks for more than the clip length can hold, or because it is full of words like "cinematic" and "stunning" that carry no actionable information. Cut it to one action and replace praise with concrete detail.
When should I use reference images instead of describing something?
Whenever a specific real thing has to be accurate — your product, a particular face, an established art direction. Text will invent something plausible; only a reference pins the actual thing.
How do I make videos longer than fifteen seconds?
Build the sequence from several generations. Video Extend continues a shot past its final frame, and Video Transition generates the movement between two shots — both work better than asking for one very long take.
What happens if a generation fails?
The credits reserved for it are returned automatically — you do not need to raise a ticket. Costs are also quoted from your exact settings before you generate, so nothing is charged as a surprise.
Keep going with the rest of the toolkit
Each tool is the same Seedance 2.0 model pointed at a different job. Move between them freely — your credits, history and exports are shared.
Generate your first Seedance 2.0 video now
Write one line, pick a ratio, press generate. No editing suite, no render farm, no post-production pass — a finished clip with sound, ready to download.
- Free credits on sign-up
- Paid commercial use
- Refunded if a generation fails