The Web SDK and TypeScript API were straightforward to inspect and integrate, but actual speech output could not be tested without the licensed voice model.
- What worked
- The installed type declarations made the worker API clear enough to build and type-check the adapter.
