Skip to content

Questions, answered plainly.

What people ask before they start, by topic. Each answer is how the app works today. Where the answer is not yet, it says so.

Asked first.

Is there a subscription?
No.Add funds and pay for what is made. Nothing renews, nothing expires.
Do I need a card to try it?
No.Sign in, finish the tutorial in the app, and $3 lands in your balance.
What does video cost?
$0.29 a secondAt 720p, to the nearest cent. $0.60 at 1080p. Every price
Am I charged when a render fails?
$0.What was held for the job goes back to your balance if it fails or is refused.
Can I use it for client work?
Yes.What you make belongs to your organization, which can be just you. Ownership and privacy
Is my work used to train AI?
Never by us.The AI companies get it under business terms that say they do not train on it. The detail
Does it make vertical video?
Yes, 9:16.And 16:9 and 4:3.
How long does a render take?
About 4 minutes.For a 5-second clip. 5 to 7 minutes for 30 seconds.
Can my team share it?
Yes.Adding people is free, and each person has a spending limit.

Starting

What do I do first?

Sign in with Google or with a code sent to your email, and confirm that you are 18 or older. A finished project, “Paradise, Found”, is already in your account, and the tutorial in the app walks you through it. Finish the tutorial and $3 is added to your balance: enough for about 10 seconds of 720p video, or for a few stills and a shorter clip.

Is it free to start, and when do I pay?

Planning is free and needs no card. The $3 from the tutorial in the app pays for your first pictures and a short clip. After that, paid work starts when you add funds, from $25 at a time.

Do I need a finished script?

No. Bring a script as a PDF, a Word file or pasted text and it is read for you. Or start with nothing: add scenes by hand, paste a shot list, or describe a scene to your own LLM and let it write the shots.

Is it for films, series or verticals?

All of them. A project is set up as a feature film, a series or a short, in a frame of 16:9, 9:16 or 4:3. Vertical dramas are a series at 9:16. Narrative work is what it is built for: the same cast, places and look from the first shot to the last.

Can I use it by myself?

Yes. An organization can be just you. One person gets everything a team gets, and adding people later costs nothing.

Does it work on a phone?

It runs in a web browser, with nothing to install. It opens on a phone, and the tutorial in the app says itself that it is best on a computer.

Can I bring my own pictures, products or talent?

Yes. Upload a picture of a product, a place or a person you have the right to use, and save it as an asset. We register real people with the video model at no charge, so it does not refuse them.

Control

How does a character stay the same across scenes, episodes and costume changes?

The face is made once and locked. Each costume is its own wardrobe sheet, drawn against that face. Every shot the character is in, in any scene or episode, is sent those pictures.

Will the video match my shot exactly?

Not every time. The app helps you frame the prompt, and a still shows you the frame before you pay for video. A video model can still stray from it. A render that finishes is charged, whatever is in it; only one that fails or is refused costs nothing. When a shot strays, you render that shot again on its own.

Can I redo one shot without paying for the whole scene again?

Yes. A still is framed one shot at a time. For video you choose which shots a render includes, so one shot can be rendered again on its own. A render is 4 seconds at the least, and it comes back as its own clip.

Which AI models does it use?

Stills: GPT Image 2, GPT Image 2.5, Seedream 5.0, Seedream 4.5, Nano Banana 2, Nano Banana Pro. Video: Seedance 2.5, Seedance 2.0, Seedance 2.0 Fast, Seedance 2.0 mini, Veo 3.1, Veo 3.1 Fast, Veo 3.1 Lite. A sequence you send to video opens on Seedance 2.5, and the engine menu there offers the others. Upscales go up to 4K. You choose the model for each picture and each render, and the price on the button changes with it. The Models page lists every one with its price.

Can I use it with Claude or ChatGPT?

Yes, and you do not have to. An LLM is an AI assistant such as Claude or ChatGPT. Connect your own, on your own account with its maker, to the projects you choose. It reads the script, writes shots and proposes pictures and renders. Every paid proposal waits for a person to press Approve, so it cannot spend your balance on its own.

The video

What exactly comes out?

On Seedance 2.5: video at 480p, 720p (HD) or 1080p (full HD), at 24 frames a second, 4 to 30 seconds a render, in 9:16 (tall, for phones), 16:9 (wide) or 4:3, as MOV or MP4, with sound in the same file and no visible watermark. 2K and 4K come from upscaling a finished clip, which changes nothing in the shot. Seedance 2.0, Veo 3.1 and Veo 3.1 Fast can also render at 4K.

How do I get my clips out?

Download the original file of each clip. Renders are downloaded as MOV or MP4 clips and cut together in your own editor. Sending takes to Frame.io and to your own Google Drive is being built.

Money

How do I pay, and can I get an invoice?

By card, $25 to $1,000 at a time, with automatic top-up if the owner authorizes it. Every payment has a receipt. There are no invoices and no separate sales contract: you accept the Terms online, add funds and use the tool.

Does my balance expire, and can I get it back?

It never expires. The owner can ask in Billing for the unused funds the organization paid for: they go back to the card that paid within 10 business days. Money spent on finished work and promotional credit are not refunded.

Can I see what each project cost?

Yes. Each project has a Budget page: what it has spent by scene, sequence, person and model, against a budget you set, with filters and a CSV. The billing history lists every job with its price and who ran it.

Team and clients

Can teammates sign in without a Google account?

Yes. Anyone can sign in with a code sent to their email. There is no password. Two-factor sign-in can be switched on, and an owner can require it for everyone.

Who can see our projects?

The people you invite. Owners and managers can open every project in the organization. Members and viewers can open the projects they are given, and their own loose work. A role decides what each person may change, generate or spend. Story Atlas staff see who is on the team, counts, money, timings and error codes, never scripts, boards, prompts or files. Two exceptions: a request our safety check blocks, which a person reviews, and a file named in a report or a legal notice, which a person may open to decide whether it must be removed. The owner can let support in for 24 hours: support then also sees the names of the projects and how many sequences each has, not what is in them. For a legal or safety case Story Atlas can give itself that same view for four hours without asking, and it is written in the security record as “Emergency access by Story Atlas”. The Privacy Notice has the rest.

What happens to our work if the service closes?

You are told in time to download it.

Who makes it?

Story Atlas, a video production company in California. We built it for our own productions and use it on them.

Open the finished project. Then make yours.

No card needed. Finish the tutorial in the app and $3 is added to your balance.