MMM-OpenAI

sdmydbr9

@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.

sdetweil

@sdmydbr9 use a GoogleAssistant recipe

see it’s recipies folder

MMRIZE

@sdmydbr9
I have a few already, like MMM-Hotword or MMM-Clap, and some…

Anyway new OpenAI’s Whisper looks so interesting, So I’ll make it as a STT module in a few days, hope so.

sdmydbr9

@sdetweil I just meant it as a similar approach, I don’t actually want to implement the whole assistant module.

sdmydbr9

@MMRIZE I think the Hotword module is not Available anymore and the clap module, might work but it will not look as asthetic as activating with the hot word…

Rags

@MMRIZE said in MMM-OpenAI:

@Rags
Speech example: Conjunction with MMM-GoogleTTS, now it can speak.

https://youtu.be/daWuSfQc5Ns
postProcessing: (helper, resp) => {
  let choice = resp.response?.choices[0]
  if (choice) helper.sendNotification('TTS_SAY', choice?.message?.content ?? choice?.text)
}

I am getting a syntax error in these lines can you please check.
Thanks

Rags

@MMRIZE
It’s possible that the error is caused by using optional chaining in a JavaScript version that doesn’t support it.

I modified the code as below, rewriting the code without using optional chaining.
postProcessing: function(helper, resp) {
let choice = resp.response ? resp.response.choices[0] : undefined;
if (choice) {
let message = choice.message ? choice.message.content.choice.text : undefined;
if (message) {
helper.sendNotification(‘TTS_SAY’, message);
}
}

I am not getting any syntax error now. However, there is some issue with Google TTS (unrelated I think) , as it is sending a notification as TTS_Say_Error. Google TTS API hopefully is configured correctly as no errors are thrown up during starting MM. How should i go about resolving it.

MMRIZE

@Rags
Give me the front and backend log on error happens

MMRIZE

@Rags
TTS_SAY notification could have more complex payload like this.

.sendNotification("TTS_SAY", {
  content: "Live long and prosper", //text to be spoken
  //belows are optional
  type: "text", // "text" or "ssml"
  voiceName: "en-US-Standard-B",
  languageCode: "en-US",
  ssmlGender: "FEMALE",
  callback: (error)=> {
    console.log(error) // <= What you need.
  }
})

Rags

@Rags said in MMM-OpenAI:

@MMRIZE @MMRIZE
Thanks for the update. Yes now it works fine. I am not getting "something is wrong’ messages, but prompts about the token lengths whenever they exceed.
Thanks. I hope there will be voice integration soon.

So once again some issue has cropped up. The Openai integration with Telegrambot was working fine up to now. However, now the /chatai messages are not getting displayed on the MM whereas they are getting displayed on the Telegrambot app on the phone. The /txtai and /imgai have no issues. Please help in getting it to work again.