Based on the previous project cow deployment, many friends have asked a question many times. I will summarize it here and give you the answer.
No QR code appears
When you deploy and the QR code does not appear, there are two reasons. One is thatthe dependency is not installed properly (there is a red color during installation) It means there is a problem with the dependency installation) The other one is the config.json file you modified in the previous one. There is a mistake in the file. Please check again
Note: You need to log in again every time you modify the config.json file.
renew
If you update, you can check this URL to check the update progress.
Releases · zhayujie/chatgpt-on-wechat · GitHub
update command
First enter the chatgpt-on-wechat directory and enter the following command to update the project
git pull
Also first enter the chatgpt-on-wechat directory
git describe --tags
There should be some partners who have this problem of not being able to keep up with new people.
If you use the server I recommend, we can solve it here
Click to open the directory and update it.
draw
Some people like MJ drawing, some people like Dall-e < a i=4>Huahua, both can be used in the cow project. Just use it according to your own needs
Note: This function requires the linkai key
vi chatgpt-on-wechat/config.jso
First open the linkai plug-in in the plug-in and change the configuration file parameters to true
vi chatgpt-on-wechat/plugins/linkai/config.json
Then open the reference created in the linkai URL Dall-e-3 Plug-in
Demo
dalle when drawing, usehelp me draw ordraw Wait for the words to express that when you want to draw a picture, he will draw it
voice
linkai has synthesized voice, but because WeChat has restrictions onitchat protocol, it can only send voicemp3 file, if you use wechaty, the reply will be WeChat voice. I haven’t researched it eitherwechaty If you are interested, I can research it
Now I will teach you how to use voice in the cow project
vi chatgpt-on-wechat/config.json
After opening, you need to add several parameters
"speech_recognition": True, # 是否开启语音识别 "group_speech_recognition": False, # 是否开启群组语音识别
"voice_reply_voice": False, # 是否使用语音回复语音,需要设置对应语音合成引擎的api key
"always_reply_voice": False, # 是否一直使用语音回复
"voice_to_text": "openai", # 语音识别引擎,支持openai,baidu,google,azure
"text_to_voice": "openai", # 语音合成引擎,支持openai,baidu,google,pytts(offline),azure,elevenlabs
"text_to_voice_model": "tts-1",
"tts_voice_id": "alloy",
Here "text_to_voice_model": "tts-1", The parameter has several options
whisper-1 1 points = 1000 token (Note: I don’t know why this parameter will output MP3 format but cannot be played. I will remind everyone if it is changed later)< /span>
tts-1 1 point = 12 characters
TTS-1-HD 1 points = 6 characters
Here & quot; tts_voice_id & quot ;: & quot; alloy & quot; a i=3>There are several options for parameters
- onyx
- nova (only this one has a female voice)
- alloy
- echo
- fable
- shimmer
After updating their linkai website, they have a fun place
After adding sounds to the application you created, you can have voice conversations on the website just like open
Demo
If you still have any questions, you can add me on WeChat: m216m7 and join the group for consultation