Build your OpenAI ChatGPT web site.
The voice recognition function defaults to using local voice recognition mode. When local voice recognition fails, it will automatically switch to using OpenAI Whisper
for recognition. It can also be set to use Only use Whisper
for recognition.
Note: Due to the Android system's characteristics, Android phone users cannot perform local voice recognition and recording services at the same time. Therefore, mobile phone users should enable Only use Whisper
mode.
- Custom OpenAI domain name, direct connection, no need to worry about key leakage;
- Your own API Key;
- All data is stored locally;
- Model selection;
- Assistant prompt can be set;
- Multiple prompts are pre-installed;
- Session history record (local storage);
- Support for setting the
temperature
parameter; - Support for "sse", which is the
stream
in the OpenAI API; - Support for automatic text playback (TTS);
- Support for voice input (ASR).
- Support for
OpenAI Whisper
recognition (default using local voice recognition)
More people are welcome to improve this prompt list.
Fork this project and then enable your GitHub Pages.