The BEST ESP32 project based on ChatGPT | IOT Projects | NextPCB

so in this video I'm going to make my own Chad GPT based Voice Assistant using esp32 seriously let me show you a quick demo what is your name I'm an AI digital assistant and do not have a name you can call me open eye what can I assist you with tell me your favorite movie I'm in Ai and I don't have personal preferences but some popular movies include the sh sh Redemption The Godfather and The Dark Knight now here some of you people must be thinking that I already made the similar project last year well you are absolutely right last year I made a complete series of four episode of making our own chat gbt based voice assistant and you all people kind of love that video series and many of you people tried to replicate that project as well but I got a lot of feedback from you regarding a lot of issues that you were facing in the hardware in the coding and also in generating the API keys and also that series was of four episodes and many of you people missed one or two episodes so they were not able to successfully make it so now to solve all those issues I came up with the version two of the project so that by watching the single video you will be easily and successfully be able to make your own chat GPD based Voice Assistant using esp32 but the condition is you need to watch the complete video don't skip any of the part and follow all the steps shown in the video properly so now let get started with the project by covering the hardware part first so for making this project you need to have all these components whose list you can easily find in the description of this video and after that you need to connect them all according to the schematic diagram now here I'll be providing this schematic diagram in my giup repository so after finalizing the schematic we designed our own custom PCB for8 and here on the PCB we added a big hole for the speaker to get easily rest fitted on it and after that we gave the PCB order to next PCB so next PCB is one of the most experienced PCB manufacturer in China with an experience of more than 15 years years and well they got an amazing tool that will help you to locate different different issues in your PCB for free of cost let me show you so on your Windows system you can download this dfm design analysis software that will give you the detailed analysis of your PCB for example here I uploaded the Gerber file of my PCB project and when I click on this button it will start analyzing my complete PCB file and generate a dfm report for me which contains all the details about it and its issues now what's the best thing which I like in it is it not just recognizes the issue but it also let us know where the issue exactly is on our PCB and what will go wrong if we don't solve that issue for example in my PCB I'm having a silk screen design error and if I click on it as you can see I got six different points on which the error is and if I click on one of the point I can see exact location on my PCB where the problem lies so here basically the issue is my silk screen is overlapping with the component which is kind of okay for me because it's just a protype project but here you you can see that if we don't solve this issue what may get wrong while manufacturing the PCB so with the help of this thing you can modify your PCB to solve this issue before giving it to manufacturer and now as the software is from next PCB I can place my PCB order with the software itself so here I upload the Gerber file of my PCB and here I placed the order for the 10 pcbs with a green color masking and I ordered it via FedEx CER service so after a couple of days I receive my order in a box and we get to see the pcbs safely packed with the vacuum Packaging and along with this we also got two PCB scales that were having a lot of information about the pads track width wire and lot other PCB parameters then we open up the packet and here is the PCB of our project this looks really premium like the sil screen on it feels really very good so that's how you can get your custom pcbs manufacturer from next pcbs I'll be leaving over the links down in the description so after getting the pcbs one by one we started shoulding all the components on end now here we just use the female Buck strap for all the components so that we can reuse them once we are done with this project now in this project we were having only one single SMD component which is this hd73 voltage regulator IC so here after shouldering all the components the final project looks like this neat and very compact so with this we are done with the hardware part of the project now let's jump on to the coding and the API section so here first of all we need to get two API Keys one from Google cloud and other from AI so now first to get the API key from Google Cloud you need to go to cloud.google.com and here just login with your Google account then here you need to search for speech to text click on this link and here we can see that we get a $300 free credit initially to use these apis so let's click on try for free here I'll select the country as India then provide my business name here and at this point you need to provide your credit card details now here after you provide your credit card details they will charge a very little amount just for verification so don't worry so after providing my card details it ask for the CVV number and then after providing the OTP this Step One is completed and in this step two we need to provide a couple of more details now previously it was not like that like providing the credit card details was enough but this time the rules are changed you need to submit the documents to verify that the card is yours and you need to apply your identity proof as well so for the card proof what I did is I prepared this document in which I have both front and back side of my credit card visible here now here I uploaded both front and back because on the front I was just having my name written on it while the credit card number was written on the back side now here you can hide the important details like the expired date the CVV number and just show the last four digits of your credit card now here for the second document I uploaded the pan card because they need an identity proof which has the name that exactly matches with the name on your credit card and after uploading both the documents I just click on the submit button and now it says to wait for few days for verification but luckily my verification was done in couple of minutes and now here I can just refresh this page which takes me to the homepage of my Google Cloud account now here I'll will show you the generic process to enable the speech to text API and for that you need to search for speech to text here and just click on the speech to text API click on enable and now here just go to credentials create credentials and select API key now here it will automatically generate one API key for you which you need to copy and paste it inside your computer as will require this key at the time of coding now with this we are done with the speech to text API from Google Cloud now the next step is to get the open AI API key and for that you need to go to platform. open.com and here just login with your Google account after that just fill off your basic details and click on agree now here if you go to the usage section section you'll get to see that initially you get a $5 credit to use their API keys for completely free of cost after that just go to the API key and here first we need to verify our phone number to get the keys so after the verification just give a name to the key here I'll give the name as chat jbd and here I got my secret key which again I need to copy and paste it somewhere in my computer because now we won't be able to see this key again now after getting the key let's just try to verify if everything is working perfectly fine with open EI or not so for that I'll open the postman application and here just click on file and select new click on HTTP request and here we will use the post request method for this URL now in the header section we will add two key value pairs one is the content type which is application for/ J and second is the authorization in which we will provide miror space our open AI API key and then in the body if I paste the previous L used body which contains the model name prompt temperature and token let's see what happens so let's click on the send button now here we are getting an error called as model has been deprecated now this is one of the common error which many of you people started facing recently and to get the solution of this problem you just need to go to this page and here as you can see that the model got shut down 4th of January and we need to replace it with this new model so I'll copy it and paste it in the body and let's try it again and this time I got complete response from open AI which means everything is working fine and likewise in future if this model also gets shut down now you know how to replace the new model and how to make it work so as we got both the API Keys now let's move on to the code part and let's see how to integrate this API case in the code so now for the code what you need to do is you need to go to my GitHub repository who link you can find the description of this video and here you need to go to esp32 chat jpd V2 repository and just click on the download Z button after downloading this unzip that folder and you'll find two folders in it one is esp2 speech to text and other is the text to speech now go to speech to text first so here inside this we have all this different kind of libraries which you need to keep in this same folder only don't just replace it or don't do anything here in this single folder you need to have all this files so this is also one of the common mistake people do they just have this single file in the folder and all these files in the other folder but yeah I'll double click on it and it will open up the Ardo ID perfect here is the sketch so here I have already provided all the necessary things to be taken care before uploading to this code or to make this code work on your system first is you need to have the esp32 boards package version 1.0.6 and for having this same version first of all you need to go to ardino preferences and first you need to paste this line of uh this particular Link in your ardino preferences after you paste this link then you need to go to sketch into no then to go to tools then into boards and then into boards manager and then here to type out as esp32 and here to install esp32 from as system and here you can select the version as 1.0.6 in my case I already installed so it is not showing up so yeah that's the first step that you need to follow second step is all the files need to be in the single folder only named as es spe2 text which I already told you and third is all your credential like s say name password Google Cloud key and the open open API key need to be entered in one single header file called as credentials. so this is something which I simplified this time so ear you need to enter credentials in different different header files but now I have created one single headle file which is credentials. here you can provide the same name and password of your Wi-Fi router so I will quickly provide that after that here you need to provide the Google speech to text API key then here you can uh choose your own language so right now I chose the the language as Indian English then here are the credentials for chat GPD so first you need to provide the chat GPD token and after that here you need to provide the open AI model so earlier the model was dng but that was not working so new model is GPD 3.5 turbo then here is the variable for temperature and for the max tokens uh that you can learn more about in the documentation of chat GPT so basically all the credentials need to be entered here and this is the only header file that you need to change rest of everything will be same as it is you don't need to do any changes so after proving all these things now you can upload this code onto your esp32 board and for that you need to select the board as esp32 do it dait V1 so let me show you go to boards then to es ardino then scroll down and here you need to select the board as do it ESP V1 after doing that you just need to select the right comport and you need to hit the upload button and this will upload the code onto your uh first esp32 which is for speech to text similarly let's move on to the next code which which is for text to speech which is way more simpler than this one so for that again I'll go to the folder which I downloaded from my GitHub repository and this time I'll go to text to speech and here as you can see there's only one single Ardo file I'll quickly open up and here I also provided couple of prerequisite to make this code work on your system first is the same one 1.0.6 es board package you need to have then you need to install the audio. library which whose link is mentioned the code itself so you just need to click here which will open up a GitHub repository and here you to click on the code and download zip button after downloading this ZIP file you need to go to sketch into include library then into M add zip library and go to the folder where you have downloaded this library and here just click on that zip file and just click on the choose button this will add the library to your Audino ID make sure you add the same Library whose link is mentioned here many of the people have installed different different audio. library and L on the issues okay so that's the second requirement and third is need to enter the credential like Wi-Fi I said name and password in the same file which is here so here I'll provide the Wi-Fi s name and that's pretty much is this is the only change that you need to do inside this code rest of everything will remain as it is and now again after select uh selecting the right board and code you can hit the upload button this will upload the code for text to space and that was the simplest method uh that you need to follow to make your project work on your side without any error okay so after uploading the code on both the esp32 board let's just power up the device and ask some questions to at gbd so let me wave my hand on proximity sensor who is your creator so when the red light is turned on the project is listening to your voice and when the green light is turned on your audio is given to the Google Cloud for converting your speech to text and once it received the text it is given to the open AI server and at that time the light will be blue AI created by a team of developers I do not have a Creator in the traditional sense and as you can see after all the processing we are getting the response on to the speaker straight from the chat GPD let me ask one more question so yeah that was my chat gbd version 2 but still I will not be selling this project and I have a couple of reasons for that first of all I particularly not convince 100% with this particular version because it is not completely battery efficient it is not working great on battery even I tried replacing the hd73 with XC 6203 IC which is also an ldo but with a higher output current but still it was not perfect so I'm planning to do some modification into the hardware part and maybe launch the version 2.1 and I'm thinking of to eliminate the battery and make it work with a 5vt adapter maybe or maybe your mobile phone charg and like powering the es directly so I'm thinking about it to make it more efficient more relax ible and then launch it on my website for selling purpose and then you can you know replicate it or purchase it and can make it work on your side as well so still I need some time to do that but yeah once I do that I'll be not publishing the complete video about it because the coding will be exactly the same and in this video I Tred to explain everything to eliminate all the coding based errors I won't be making the complete video rather I'll be making some announcement or maybe YouTube short about 2.1 version so make sure you stay connected with me on YouTube and Instagram other than that if you already purchased the esp32 chat GPT kit on my from my website well you can try out my new code it will work in that kit as well and in the new code I hope you won't get any error just follow the video very very carefully so yeah that was all about this project now you let me know your suggestion about how was this project and what should I do more Improvement if I ever make version three of this project should I move to Raspberry Pi rather than using two esp32s what do you think and to share your thought suggestion down the comment of the video and also share this video with the one who are interested in making some interesting esp32 based iot project well this is definitely a great project to try out so yeah that was all from my side and now you just wait for my next video Until Then explore learn share with me Tech [Music] SMS

Channel: techiesms

Views: 75,895

Keywords: Chatgpt, ESP32, arduino project, chatgpt voice assistant, electronics, electronics projects, esp32 chatgpt, esp32 chatgpt voice assistant, esp32 projects, i2s, iot, iot projects, techiesms

Id: gGcskjKtArU

Length: 16min 27sec (987 seconds)

Published: Sat Jan 20 2024