What's changed
-
🛜 Switch to use the next server for TTS requests
- The "openai like" api request now happens on the next server, wherever you are running it, not the client, so the base URL is relative to the next server location
- This means you can use the same base URL (i.e.
http://localhost:8880/v1) for all browsers/devices (set .env to propagate URL to all) - You may need to edit your TTS base URL to be relative the server, you may not need to
-
🐛 Huge bug fixes and improvements
- Huge fix for iOS playback (doesn't stop on page turns anymore 🙌 🎉)
- Fix where the TTS would re-call the API even if preloading, if preload was not ready when cursor reached it (another huge fix for slower TTS endpoints)
Note: I am still testing for new bugs that might be introduced with switching to server tts.