Tenant caps
What per-tenant AI usage limits are, what the defaults are, and how to change them through the Coassemble API.
If your account has multiple tenants, each tenant gets its own monthly figure for AI usage, so you can divide your account allowance across them deliberately rather than leaving it first come, first served. Every new tenant starts with 250,000 narration tokens and 150 image generation events per month, and you can change any tenant's figure through the Coassemble API.
Why per-tenant limits exist
Tenants under your account share your account's monthly allowance for narration tokens and image generation events. The per-tenant limit is how you decide each tenant's share of that pool, and how you see whether one tenant is set up to consume far more than the rest.
What the defaults are
Every new tenant starts with:
- 250,000 narration tokens per month
- 150 image generation events per month
One narration token is one character of generated narration. One image generation event produces four pictures. Both are counted at the moment of generation, whether or not you publish or keep the result.
How tenant limits relate to your account allowance
The two layers are tracked independently. Your account has a total monthly allowance for each AI metric, and each tenant has its own figure underneath it. All tenants draw from the same account pool, so the sum of your tenant figures can exceed what your account actually has.
Both the account allowance and every tenant figure reset at the start of the next calendar month.
How to set or change a tenant's limit
You manage tenant limits through the Coassemble API. You can:
- Set a different monthly figure for a tenant.
- Set a tenant to "unlimited" so it draws freely from your account-level allowance.
Only the parent account can change these. Tenants can't change their own, and visibility sits at the parent account level. For the endpoint and request shape, see the developer documentation.
Seeing per-tenant usage
The Per-Tenant Usage Reporting API returns consumption broken down by tenant for any time window, so you can build your own dashboards or alerts on top of it. The endpoint shape is in the developer documentation.
This is the reliable way to keep track of distribution across tenants, and it's worth setting up an alert rather than checking at the end of the month.
Frequently asked questions
Do Identified Recipients and Anonymous Recipients have tenant-level limits?
No. Identified Recipient (IR) and Anonymous Recipient (AR) counts are handled at the account level only, and they're measured per year rather than per month. Per-tenant limits apply to AI usage, meaning narration tokens and image generation events, which are both measured per month.
Can I set limits for all my tenants in one call?
There's no bulk-set option, so you set each tenant's figure with its own API call.
Can I top up a tenant that's running low mid-month?
No. Top-ups don't apply at the tenant level. If a tenant is running low before the month rolls over, raise its figure or set it to unlimited instead.
What happens when the account allowance runs out?
AI allowances are hard capped at the account level. Once your account's narration tokens or image generation events are used up for the month, generation stops for every tenant until the allowance resets, whatever their individual figures say.