Image to video
Composer
Confirm the steps and estimated cost before generating.
Try these examples, or swap in your own media.
Keep the subject, replace a messy backdrop.
Needs: a product photo
Animate a still into a short clip.
Needs: a photo
Upload your portrait and see them talk in seconds.
Needs: a portrait (narration can come next)
Grow a usable hero image from text.
No media required — just an intention
Describe what you want to make — you can also drop in images, video, or audio.
Image to video, vertical product motion and lip-synced presenters. Pick one to load the same request into Composer.
A close-up reference turned into a more tactile motion clip.
Start from thisA natural portrait scene animated into a lighter social-style result.
Start from thisFrom one sentence to image, video, audio, and talking avatars. You can always open a studio directly instead.
Agents
Studios
AI quota covers understanding and planning. Coins cover actual generation. You'll see the estimate before anything runs.
AI quota
Composer reading your material and Campaign planning a kit both draw on AI quota, measured in tokens.
Free includes 50,000 tokens a month. Ultra includes 30,000,000.
Coins
Image, video, audio and lipsync generation spend coins. Rates vary by model and duration.
Free includes 100 coins a month. Watermark removal is 2 coins per clip.
Exchange
Running low on AI quota mid-month? Convert coins at 1 coin = 1,000 tokens, any time.
One-way: quota does not convert back into coins. Both reset monthly.
Every plan includes both meters: coins for generation, AI quota for understanding.
$0
$9.9/mo
$39.9/mo
$79.9/mo
Didn't find your answer? Contact us any time.
Composer
Composer is the default way in. Describe the outcome in one sentence and drop in any material. Composer reads what you gave it, returns a plan card with the goal, the steps, which materials it will use and a coin estimate, then runs it once you confirm. You keep editing the same piece instead of hopping between tools.
Studios hand you the raw parameters for one capability. Use Composer when you are describing an outcome; open a Studio when you already know the exact model, duration and aspect ratio you want.
Composer works on a single piece and its revisions. Campaign starts from a kit — ads, commerce or a social pack — and walks you through a fixed sequence of steps until the whole set is finished. Use it when you need several matching assets rather than one.
Credits & plans
Two separate meters. AI quota, measured in tokens, covers understanding and planning: Composer reading your material, Campaign planning a kit. Coins cover generation: image, video, audio and lipsync. The coin estimate always appears on the plan card before a job starts.
100 coins and 50,000 AI tokens a month, one concurrent task, up to 60 seconds per avatar video, and watermarked downloads. Paid plans start at $9.9/mo with 300 coins, 300,000 tokens and watermark-free exports.
Yes. Convert coins into quota at 1 coin = 1,000 tokens, any time. The conversion is one-way — quota does not turn back into coins — and both balances are granted monthly.
Capabilities
Video: LTX, Veo 3.1 and Seedance 2.0. Image: Nano Banana Pro, GPT Image 2 and Midjourney. Plus lipsync, text to speech, voice cloning and stem separation. Seedance 2.0 is included on every paid plan.
Usually one to five minutes, depending on length and queue depth. Watermark removal finishes in under five seconds. Paid plans get more concurrent slots, from 2 on Starter to 50 on Ultra.
9:16, 1:1 and 16:9, at bitrates suited to social platforms. Every paid plan exports without a watermark; free downloads carry one.
Rights & privacy
Paid plans include commercial use. You are responsible for holding the rights to whatever you upload, including likeness and copyright, and watermark removal is only for material you own.
They are used to run your jobs, never shown publicly, and you can delete them at any time. See the privacy policy for the full detail.
The interface is available in English, 简体中文 and Español. Narration and voice cloning cover a wider set of languages, with more being added.