Free
Explore a complete Wok before committing.
$0forever
- 1 live app with linked Preview
- 1 concurrent hosted Wang run
- DeepSeek Flash for Wang app work
- 1M model tokens and 1M requests monthly
- Prepaid usage after included allowances
Pricing
Pay for the Wok.
Keep usage honest.
Every plan includes the complete stack, a linked Preview, and a hosted Wang Agent Cell. Usage beyond included allowances is prepaid while postpaid usage billing is disabled. Choose by the number of live apps, Wang concurrency, model access, and the recovery shape the workload needs.
Pro
Ship one serious product with Wang alongside you.
$49/ month
- 1 live app with linked Preview
- 1 concurrent hosted Wang run
- 10M model tokens and 1M requests included monthly
- Prepaid usage after included allowances
- Custom domains and portable exports
Scale
Run several products and parallel agent work.
$149/ month
- 5 live apps, each with a linked Preview
- 4 concurrent hosted Wang runs
- 10M model tokens and 1M requests included monthly
- Prepaid usage after included allowances
- Custom domains and portable exports
Scale + HA
Add failover where one database process is not enough.
$199/ month
- Everything in Scale
- 10M model tokens and 1M requests included monthly
- Prepaid usage after included allowances
- Database standby and failover
- Redundant REST service
- Recovery design set around the workload
Enterprise
A contracted Wok fleet with deeper isolation.
from$499/ month
- 10 live apps to start
- 8 concurrent hosted Wang runs
- 10M model tokens and 1M requests included monthly
- Contracted usage billing
- Dedicated or confidential isolation
- Contracted limits, recovery objectives, and support
Compare
The plan changes capacity, not the shape of your stack.
Postgres, Auth, Storage, Realtime, Functions, Scheduler, Email, hosting, and the model gateway remain one Wok.
| Included | Free | Pro | Scale | Scale + HA | Enterprise |
|---|---|---|---|---|---|
| Live apps | 1 | 1 | 5 | 5 | 10 to start |
| Linked Preview | Included | Included | Included | Included | Included |
| Included model tokens | 1M | 10M | 10M | 10M | 10M |
| Bundled storage | None | None | None | None | None by default |
| Concurrent Wang runs | 1 | 1 | 4 | 4 | 8 |
| Hosted Agent Cell | Included | Included | Included | Included | Included |
| DeepSeek in Wang | Flash | Flash + Pro | Flash + Pro | Flash + Pro | Flash + Pro |
| Usage funding | Prepaid | Prepaid | Prepaid | Prepaid | Contracted |
| Kimi model | — | Included | Included | Included | Included |
| Custom domains | — | Included | Included | Included | Included |
| Database standby | — | — | — | Included | Included |
| Confidential isolation | — | — | — | — | Available |
Usage
Published usage rates.
The base plan covers the operating shape. Usage beyond included allowances uses prepaid credits while postpaid usage billing is disabled. Existing postpaid charges remain attached to the subscription that incurred them and continue through normal reconciliation. Storage keeps its published rate while collection waits for a pinned meter and trusted quota. Planned meters are labeled.
Private per 1M: $0.001552 cached input, $0.18618 input, $0.37236 output. Kimi per 1M: $0.2033 cached input, $1.0165 input, $4.28 output. Usage beyond the included model allowance is prepaid.
Planned: 100k included, then $2 per million on Free and $0.80 per million on paid plans.
Prepaid. Carrier and registration charges can still apply.
Collection rolls out after the meter is pinned and the quota is trusted.
Each recipient counts as one send.
The current debit is collected when the number is provisioned.
Flux Pro is $0.12 per image. Recraft SVG is $0.40 per image.
Charged when a try-on request is accepted for processing.
The exact registration and renewal price is shown before purchase.
Questions
The details that affect the bill.
What does Scale + HA include?
Scale + HA adds a database standby and redundant REST service. Placement, recovery objectives, and failover design are set around the workload.
Does a linked Preview count as another app?
No. A live app and its linked Preview consume one app slot together.
Why separate the plan from usage?
The subscription covers the persistent stack and operating capacity. Metered services use the published rates above, so a light-use team does not subsidize an unusually heavy workload.
How is metered usage paid?
Included grants are used first. Usage beyond those allowances uses prepaid credits at the published rates. Postpaid usage billing is temporarily disabled. Storage collection waits for a pinned meter and trusted quota.
Can we move a Wok elsewhere?
Yes. A Wok uses standard Postgres and its definition is exportable, so paid plans can redeploy it on another cloud or a self-hosted fleet.
Ready when you are
Start with the app. Keep the whole stack.
The free plan is enough to build and verify a complete Wok.