- Use gpt-4-turbo-preview rather than gpt-3.5-turbo-instruct. Much smarter and wittier
- Use OpenAI streaming API to make the speech-to-response time even faster than before despite gpt-4 being slower
- With OpenAI, your chats are now saved (up to 16 messages) and used in the context of each request. So the model can recall previous requests
- Increase token expiration from one day to one month
This release is untested. Let me know how it goes
Targets
- Windows (10/11, maybe 7)
- WirePodInstaller-v1.1.3.exe
- macOS (11 and above)
- WirePod-v1.1.3.dmg
- Android (4.4 and above)
- WirePod-1.1.3.apk