Start free. Pay when it's paid work.
Free is non-commercial rather than crippled — you get the whole engine. The moment you're doing paid work, you're on Individual.
Three rungs, and one of them is free.
The whole engine, for learning and non-commercial work.
- All ten node families
- The built-in assistant
- Ship scenes as a link
- Community Recipes
- Hosted generation, capped
For solo creators doing commercial work.
- Everything in Free
- Commercial licence
- Bigger generation bundle
- Unbranded exports
- Priority queue
For teams with operators, brand systems and IP constraints.
- Author and operator seats
- Shared Recipes and workspaces
- White-labelled deliverables
- Run inference on your own hardware
- Contractual no-training guarantee
Early access, stated plainly: Recipes ship with the beta rather than today, and running inference on your own hardware is an offering we are defining with the first teams that need it. Everything else listed above is live now.
A subscription, plus what you generate.
The engine is the subscription. Generation costs real money to run, so it's metered — the way it is everywhere else in this category. We'd rather explain the mechanics than hide them behind a round number.
Every plan carries an allowance
Each plan includes a monthly amount of hosted generation. Most work sits inside it — you only think about this if you're generating at volume.
Different models cost different amounts
A quick draft pass and a slow high-resolution one are not the same amount of compute, so they don't draw the same amount from your allowance. The rate is shown before you run.
Heavy users can add more
If a project eats the month, add generation without changing plans. And on Studio and Enterprise you can run models on your own hardware instead, where the metering question goes away.
We're not going to print an image count here that turns out to be wrong. The allowance for each plan and the current rate for each model are published in the app, and both are being tuned during early access — as model costs move, so do they. If a number matters to your budget, ask us and we'll tell you what it is today.
Whose name is on the work.
The rungs aren't really about how many features you get. Each one buys a deeper claim on how the work arrives at the person who sees it.
The work carries our shell
Scenes you share open in a nooodles share shell. It's how a free tier pays for itself — the work is yours, and the page it arrives on says where it was made.
Our branding comes off your exports
What you send looks like your work, not like a tool's output. This is the line most solo commercial work needs, and it's the first thing $19 buys.
The deliverable is white-labelled
The thing your client receives — the scene, the operator surface your producer runs, the link you hand over — carries your studio, not us.
The tool itself is white-labelled
Editor and runtime under your own name, alongside the pieces that only matter at this size: a contractual no-training guarantee, and inference on your own hardware.
The things people actually ask.
What counts as commercial?
Work you're paid for, or work made for a business. Client jobs, freelance work, anything shipped by a company, anything on a product you sell. Learning, personal projects, coursework, experiments and portfolio pieces that aren't part of a paid engagement are fine on Free. If you're not sure, the honest test is whether an invoice exists.
What happens when I run out of generation?
Generation stops; nothing else does. The editor keeps working, your graphs keep cooking, your scenes stay live, and everything that isn't hosted inference carries on as normal — which is most of the engine. You can top up, wait for the month to roll, or move that work to your own hardware on Studio and Enterprise.
Do unused amounts roll over?
We're finalising this during early access, so we're not going to write a policy here that we then change. Tell us how your work is shaped — steady, or spiky around deliveries — because that's exactly the input we're using to decide.
Can I self-host?
Generation runs where you choose — hosted by default, and on your own hardware when the work can't leave the building. That's the Studio and Enterprise option, and the managed version of it is being defined with early customers rather than sold off a page. See motion design studios for what that's for.
Do you train on my work?
On the hosted tiers we may use graph structure — how nodes are wired, not your source material — to improve the built-in assistant, and that's disclosed in the terms. A contractual no-training guarantee is available for enterprise. On a local or self-hosted setup the question is settled by construction: nothing phones home.
Can I use the output commercially?
Yes, from Individual up — that's what the commercial licence is. Free is for non-commercial work. Where generative models are involved, the terms of the model you generated with also apply, and we surface those rather than paper over them. Worth knowing: because a nooodles graph is an explicit set of human decisions, AI-assisted work made this way holds the strongest authorship position such work can hold. That isn't a guarantee of ownership, and anyone telling you otherwise is selling something.
Do you offer education pricing?
Yes — nooodles is free for students and teachers, and there's a route for classrooms and labs that need more than the individual free tier. Details on the education page.
How do I cancel?
From your account settings, in a couple of clicks. No call, no retention script. The plan runs to the end of the period you've paid for and then stops. Your files stay yours — projects and scenes are documents you can save locally, and anything you've already exported is unaffected.
Open it and see what it is.
Free to start, nothing to install. Or tell us about your team — seats, operators, IP constraints — and we'll put a number against it.
Working with a team?
Tell us where to reach you and we'll set up a walkthrough.
No spam. A real reply from a human.