Common Questions
Answered

Billing follows two things: the languages you switch on, and the number of distinct endpoints in the schema you publish. Switching on a language lets you generate as many clients in it as you like, but every client built from that language draws on one shared endpoint pool. Endpoints are counted by the unique operation identifiers in your schema, never by the number of clients you ship. On the Studio plan that pool is capped at 250 per language, so you can publish an internal client and a partner client side by side and stay inside the limit, provided the two together reference no more than 250 distinct endpoints.

One method paired with one path, as declared in your schema. Two verbs on the same path count as two, while the same endpoint reused across three clients still counts once.

No. Anything filtered out before generation is left out of the tally as well, and the build log prints the exact set that was counted so you can reconcile it against your own inventory.

Builds keep running. You get a notice at 90 per cent of the pool and a prorated quote for the next tier up, which you can accept whenever it suits your finance cycle.

You do. Every client we generate lands in your own repositories under a permissive licence, so the packages your consumers already depend on keep building long after the subscription ends.