Pricing
You are paying for machine time, not for features.
Every tier is the same application. What changes is how much CPU and GPU session time you get, whether your work survives the end of a session, and how many copilot tokens are included.
Every price and limit on this page is a placeholder. Chudware is not accepting payment yet, and no figure here is an offer. Real numbers go up before billing is switched on.
Tiers
-
Evaluation
$0 / no card
Placeholder limits
For deciding whether the thing works for you at all.
- CPU seat. Software rendering, shared with other evaluation sessions.
- No persistent session. The container and everything in it is discarded when you disconnect — export before you leave.
- Idle sessions are reclaimed after 30 minutes without traffic.
- Copilot tokens: TBD per month.
- No PLM project vaults, no desktop build.
-
Individual
$49 / month
Placeholder price
For one engineer doing real work.
- Persistent home directory. Your files survive session restarts.
- One seat. Current session defaults: 2 vCPU, 5 GiB RAM.
- GPU-accelerated seat when one is free, with hardware video encode.
- PLM project vaults and the parts library.
- Desktop build for Windows and Linux.
- Copilot tokens: TBD per month, overage TBD.
-
Team
$79 / seat / month
Placeholder price
For a group that shares projects and needs someone in charge of it.
- Everything in Individual, per seat.
- Shared PLM projects with per-project history across the team.
- Admin console for approving accounts, assigning seats and revoking access.
- Priority on GPU seats when they are contended.
- Pooled copilot tokens: TBD per seat per month.
- Minimum TBD seats.
What the price is actually made of
Chudware sessions are not cheap to run, and pretending otherwise would just produce a pricing page that has to change later. These are the four things that cost money.
-
CPU seat time
A default seat renders and encodes in software on shared host CPU. It is the floor cost of having a session open at all, which is why the free tier does not keep one warm.
-
GPU seat time
Accelerated seats get a slice of a GPU for rendering and hardware video encode. There are only so many at once, so they are metered and prioritised rather than unlimited.
-
Copilot tokens
The AI copilot calls a language model, and that is billed per token upstream. Included allowances differ by tier; heavy use is metered instead of throttled silently.
-
Persistent storage
A home directory that survives disconnects, plus project vault history, is real disk that has to be kept and backed up. The evaluation tier skips it entirely.
Billing is not live yet.
Access today is granted by approval. When checkout opens, an active subscription will be one of the ways an account is cleared.