Create a Smart Voice Assistant using Open AI's ChatGPT API, Whisper, Python & Gradio

OpenAI launches APIs for ChatGPT and Whisper. In this video, I'll show you how you can create your own Smart Virtual Assistant like Google Assistant using Open AI's ChatGPT API, Whisper. The entire solution will be created using Python & Gradio. The API is priced at $0.002 per 1,000 tokens (~750 words), or 10X less expensive than any of OpenAI's previous GPT-3.5 models. After some extensive optimizations, using ChatGPT costs 90% less than it did in December. This means that the gtp-3.5-turbo model (which powers the ChatGPT you may have used) now costs $0.002 per 1,000 tokens. To put it another way, you can get 500K tokens for a dollar (ChatGPT doesn’t work on single characters, a token is more similar to a vowel). This new price makes it much cheaper for companies to build their own chat bot. But ChatGPT works with text, what if a company wants its bot to answer to voice commands? That’s where Whisper comes in – this model can transcribe speech into text at a cost of $0.006 a minute (i.e. $1 for 2 hours 45 minutes). Not only does Whisper understand 99 languages, it can also provide an English translation of what was said. This has a lot of potential to disrupt Siri and Google Assistant, which can only handle a few languages. Companies can use OpenAI’s APIs to build new chat bots and some already have, Microsoft isn’t alone.

*I use affiliate links on the products that I recommend. These give me a small portion of the sales price at no cost to you. I appreciate the proceeds and they help me to improve my channel!

Equipments I use for recording the videos:

If you do have any questions with what we covered in this video then feel free to ask in the comment section below & I'll do my best to answer those.

If you enjoy these tutorials & would like to support them then the easiest way is to simply like the video & give it a thumbs up & also it's a huge help to share these videos with anyone who you think would find them useful.

Please consider clicking the SUBSCRIBE button to be notified for future videos & thank you all for watching.

You can find me on:

#chatgpt #api #openai

[asadanauyo]

Is it possible to use like Alexa? With waking word? :)

[gameboyfrank]

This is so amazing! I have been looking for adding voice to reply from chatbot all the time. And your video is amazing!!

[sudhakardubey]

@BHAVESH KUMAR I am getting this Error
The string did not match the expected pattern

[kennethbeal]

Nice, thank you! When I left "large software company, " I realized I could do "almost everything" that was being done there; just, a bit slower, and with less "infrastructural support." Was really neat though. This seems similar! And, it could be that Alexa/Echo/etc incorporates something like this into their solution(s). Who knows, they might even hire you! :)

I bought a MyCroft, couldn't activate it (the rep agreed it was an issue with their website), and now it's a few months later and the company's on their way out. Perhaps I'll be able to integrate something like what you're demonstrating, so the purchase isn't a complete waste? (Also I need to look further into how the company is handling their demise...)

[TanmayShrimali]

Hello tried your method and same method with other person website
Yet my output fails on Right side in gradio interface
non coder here!
Whats wrong?
Recording seems to give a correct playback but no output.
plus correct api key inserted also from open ai and also my open ai credits are still available!

[amirulmukminin]

help me how you get the GPT_secret_key.json file?

[AlexanderMorou]

Quite a few changes would need made for a scalable version for an assistant triggered by keywords, but it is a neat demo.

[anjalijain]

Not working showing unexpected end of JSON output

[tradingwithwill]

my gradio interface didn't open. I got this to work using chatgpt help you have to put share=True in the launch function

[tradingwithwill]

helpful. just an fyi now you're using open api key instead of session (which is a separate payment). also the session for me didn't work it tries to open a browser in collab and issues around the selenium chromium libraries.

[jaeson]

Getting ffmpeg command line failures. Need some explanation on that line.

[Zovi-kanu]

from where we get gpt secrets key json file

[jamhulk]

How about after the reply it will wait for your next instructions/questions

[CelalEksi-jxdh]

Hello, I have a project homework but I couldn't do it, can you help me?
I need to make a question-answer program using the Openai library that asks questions with voice, provides text-based answers.

[musumo]

Thanks! This is awesome. What/how do you recommend to create a frontend? Or is there a way to turn this into an API? I’m Very new to coding thx

[reelsgalataa]

Hii bro.. 2 days before I run the code and get the output but today it shows unexpected end of Json input ERROR. Can you resolve it, and tell us what is the problem and how to resolve

[k.spremkumar]

Unexpected end of json file error. please help me on this

[calfisionxyz]

hello bro in the line openai.api_key=["API_KEY"] the error was showing like "string indices must be integers " how to resolve it plz bro tell me fast i have project

[harshsindwani]

can you make a new video of speech to image generation using open ai API, i have tried the previous video and the process is really very slow
is there any method to speed up the process

[gameboyfrank]

there is this line of code "!ffmpeg -f lavfi -i anullsrc=r=44100:cl=mono -t 10 -q:a 9 -acodec libmp3lame Temp.mp3" that I runned in python but it didn't work out. Can you help?