Read the statement by Michael Teeuw here.
MMM-OpenAI
-
MMM-OpenAI
MagicMirror module for requesting OpenAI’s API
DISCLAIMERS :
This is not aThis module uses OpenAI’s API for text completion and image generation.ChatGPT
module; (OpenAI has not released ChatGPT API yet.).- I will not implement any
voice-related
features into this module directly, but if you want, you can relay those kinds of modules for your own purpose. (This module has much preparation to cooperate with other modules.)
Screenshot
Features
- Request text completions or image generation from OpenAI with detailed options
- Activation through TelegramBot (default behaviour)
- Activation with notification (customizable - For developing a new module or For using the current module without modification, either possible.)
- Customizable post-process (You might not need to build an external module to handle the response for your own purpose)
For more details
-
1.1.0 (2023-03-02)
Added : Chat API
-
Great. Installed it and it is up and running with absolutely no effort. Thanks for this module. However, i am facing an issue of an error message “Something is wrong see the log”. So this error pops up randomly when i initiate a /chatai or a /txtai or /imgai, and i am unsure why. Is the heavy traffic of ChatGPT the reason. Also where do i check the logs.?
Thanks -
-
@Rags
I’ll update the modules with more kind error messages in 1-2days. Plz wait a little -
1.2.0 (2022-03-06)
- ADDED : More kind error message (in the terminal console, the display, telegram, and the dev-console)
- ADDED : Token usage on success in TEXT and CHAT.
- FIXED : Reducing default
max_token
for chat. (Too many token limit may cause the error unexpected)
-
@Rags
I found it was token usage issue. I think I fixed it with the new update. -
-
@Rags
I will not implement Voice-feature into this module directly by myself.
But I believe other existing voice-related modules could be used together already. -
@Rags
Speech example: Conjunction with MMM-GoogleTTS, now it can speak.postProcessing: (helper, resp) => { let choice = resp.response?.choices[0] if (choice) helper.sendNotification('TTS_SAY', choice?.message?.content ?? choice?.text) }
-
@MMRIZE a similar idea has crossed my mind. Using a Hotword detector like the one used in google assignment module to activate a listener for voice and ask for voice input and then using the input provided, get the transcript of the audio using the whisper api or something similar to get the transcript… pass on the transcript as a content in the OPENAI_REQUEST notification and then SHOW_ALERT or just read it out loud using the MMM-GoogleTTS. Just a rough sketch. I’m sure someone out here will come up with more convincing and better ideas.