ESP32 Voice Assistant with ChatGPT: Your Personal AI Companion
This is the fourth episode of our ChatGPT ESP32 series in which we made a complete voice assistant using ESP32 board in which you can ask any question to it and it will respond via speaker.
In the backend we are using usign Google APIs and Open AI APIs to make it work. Watch the video out to learn more....
⚡️⚡️ Order all components required for making this project using this single link - techiesms.com/product/chatgpt...
To try out free trial version of Altium, click here - www.altium.com/yt/techiesms
Get the best quality components from Octopart:
octopart.com/
You can support me🙏🏻 and also get benefitted on Patreon :- / techiesms
__________________________________________________________
👨🏻💻 Important Links 👨🏻💻
Previous Episodes of this series
Episode 1 - • Running ChatGPT on ESP...
Episode 2 - • Running ChatGPT on ESP...
Episode 3 - • Speech To Text using E...
Code used in this project - github.com/techiesms/ESP32-Ch...
__________________________________________________________
Buy the components for your projects and support techiesms💡🔌
⚡️⚡️ Order all components required for making this project using this single link - techiesms.com/product/chatgpt...
Or if you want to order only specific components then here is the list and link,
ESP32 Dev board - www.techiesms.com/product/esp...
I2S Mic - www.techiesms.com/product/inm...
I2S Audio Amplifier - www.techiesms.com/product/max...
Speaker - www.techiesms.com/product/rec...
IR Sensor Module - www.techiesms.com/product/inf...
TP4056 Battery Charging Module - www.techiesms.com/product/tp4...
Battery - www.techiesms.com/product/kp-...
and many more such components are available at www.techiesms.com
__________________________________________________________
⏱ Time Stamps ⏱
__________________________________________________________
🎥 🎙 My Video Gear 🎙🎥
Camera - Canon 750d
Lens - 50mm f/1.8 STM Lens (amzn.to/32gqnsi), 24mm f/2.6
Key Light - Godox SL60W (amzn.to/3522mHe)
Fill Light/Back Light - Osaka Lightning Setup (amzn.to/2I5cdDi)
Tripod - amzn.to/3uNhC4H
Microphone - Zoom H1N(amzn.to/3Osh12A)
Editing - Final Cut Pro X on MacBook Pro ( amzn.to/3vHcSyJ )
__________________________________________________________
Get connected with techiesms everywhere..
Subscribe Main Channel :- bit.ly/techiesms_youtube
Subscribe Shorts Channel :- bit.ly/techiesms_shorts
Facebook :- / techiesms
Twitter :- / imtechiesms
Instagram :- / techiesms
Website :- www.techiesms.com
Telegram - @techiesms
__________________________________________________________
techiesms
explore | learn | share
Some limitations of this project are: 1. The recording time is only 1.9 seconds. To make this time longer, you would have to increase the wavdatasize. But that wouldn't be possible since the esp32 has limited memory. However, you could use a micro sd to store the wav file and then send it to cloud to be converted. 2. Alot of times, the esp32s computational power is not enough to keep up with the audio playback. What this means is that, if you generate a response of more than 20-30 tokens, a lot of times what happens is that the esp32 doesn't synthesis the text and there isn't an audio playback on the speakers even though you have received the generated output. 3. Doesn't have a wake word detection. 4. A huge delay starting from the input to the output. A simple solution maybe to use a faster board like nvidia jetson nano which is an ideal development board with computational power more than enough for this kind of projects.
Hello I found your reply helpful, I'm inquiring about how I can be able to incorporate a wake word in this same project. Do I have to train and deploy a separate model to detect the wake word or is there a way around this?
@@nicksonamwine8869 did you find out how you can implement that?
Does this project run on esp8266?
@@olahroland3788 yeah I did, though I implemented it using python
yeah it should work with esp8266, but as he said you will have an memory issue @@atiradordeteia-brasil2752
Como siempre techiesms trayendo ideas fuera de lo ordinario. Simplemente el mejor!!!
bu video pekçok şeyi degiştirebilir . teşekkürler emeginiz için harika bir konu olmuş. artık daha akıllı cihazlar tasarlanabilir
seriously awesome bro you are doing great
This is so cool! I will build it tomorrow, thanks a lot for the great work, Edit: just finished the video, now I understand why I was having a conflict running speaker and mic !
You could try renaming one of the libraries (.h, .cpp files).
Great that you've built it. Actually I want to use this as a part of my project and it's the first time me and my group is building something like this. Ik it'd be coming off a bit wierd but is there a way we could contact u to help us out in our project if we ever need help or get stuck somewhere and need a little guidance. It'd be really appreciated. Please 🥺
Can you please share the GitHub repo of your project. We are trying to build it as our college project but we are getting an error in the text to speech part .
@@aghaaahshiiiiiiikeiji6726 hey did you built it then. if Yes . did you use a wake word instead of this IR bullshit?
Bro this is incredible 👏
🎯 Key Takeaways for quick navigation: 00:00 🎯 The video introduces a voice assistant project based on ChatGPT running on an ESP32 board. 00:26 📚 To understand this project, viewers need to watch three previous episodes in the series for the context. 01:25 🏭 The video is sponsored by Altium 365, an electronics product design platform that offers PCB design, data management, collaboration, and more. 02:08 🎤 The previous episode successfully converted speech to text using Google Cloud services, leading to filtered text responses. 03:16 🔄 Merging the speech-to-text and text-to-speech codes caused compiling errors, resolved by using separate ESP boards for each function. 03:58 🔌 The communication between the two ESP boards is done via UART, enabling a seamless flow of data for the voice assistant. 06:00 📐 After assembling the components and fitting them inside a plastic casing, the voice assistant project is ready for testing. 06:18 🛠️ Turning on the voice assistant, it responds to questions with different LED indications for recording and processing. 07:02 🗣️ The voice assistant successfully answers questions, including multilingual responses, demonstrating its functionality. 07:45 💡 Viewers are encouraged to suggest improvements for the voice assistant project in the comments section for possible future episodes. Made with HARPA AI 👍 Upvote to improve video surfing
Can you give me the name of the components used
Check description for that
@@techiesms thanks, I got it
this is awesome ,, but did you try to increase the input sampling rate in order to increase the speed of the Google Cloud API ?
Very useful project!! Now it will be possible make lots of industrial as well as educational projects .=Thanks!!
Happy Making 😇
Thank you so much, i really love your project🤖⭐⭐⭐
Now make it control electric components
For that you would want a local model
Love this Project
was that a L7805 voltage regulator used in the circuit diagram?
Perfectoo. I need ones
very cool bro... simple is beauty 👍👍👍
Thanks 👍
Professional Project Bro... 😍♥️🔥
Bro really great work ❤ can you please make this using raspberry pi
Cool next gen ESP
Hi ! I bought two BAOFENG handys, how can I transmit from one arduino to another using them ? I mean turn audio and mic plug to TX RX, from arduino to HANDY KENWOOD PLUGS (one is 3.5 stereo the other 2.5 stereo) and receive the message in another handy which receivesd the signsal and sends via plugs to ardunio and read the message in serial moriunor a
Amazing project 👍👍👍
Glad you like it
Is it necessary to connect the microphone via SPI? Can't we connect a microphone like Max9814?
Awesome 👍
I can't find the microphone and amplifier board. Could you suggest alternatives?
i can use the bard ai apk or token in place of chat gpi api directly if not what changes i need to do in my project to use the google bard api insted of chat gpt
Nice video, thanks :)
Glad you liked ✌️
Hi Sir, Can i use 1 NodeMCU for microphone, and Other ESP32 Wrooms board for Chatgpt etc?
Amazing sir can we use this project on old android phone which not support any assistant like android 2.3
You did it 🎉🎉🎉
Yessss
How many languages does this program support? Can Vietnamese be used? Thank you for answering my question. Thank you for sharing what you shared. It is wonderful.
Hi sir can you install Alexa in means we can also use homeautomation and voice assistant please reply
Hi bro loved❤ your video, but can you tell how much is the battery's voltage
Nice PROJECT
Great video 👍
Thanks 😊
Nice video,can't help about coding but for sure it needs to be fast and instead of ir sensor u can use single words same like Alexa Google , also using 2 board isn't a good idea I think someone who knows how to edit Library can change name of the same library,also isn't there any alternative for that lib
Thanks for the tip
hey, great project. In one of your previous video you said this project could be performed using esp8266 as well. It would be great if you create a video on that.
Hello , I'm working with s3 module, but i2s Library is not matched with the previous esp version . Can you help me with this issue?
I initially started watching this entire series with a hope to find a solution of problem im facing in one of my project. What im trying to do is record video and audio using esp32cam. I thought the part of speech to text might be helpful but still it will have the problem of syncing with the video. If you have done any such project earlier please point me to the link and if you haven't done it yet, then please make one project covering this part
Have u solved the solution
how did u replace 7333-A voltage regulator using a toggle switch
It would be more interesting (and usefull) if there was a way to merge this with another ESP32 that could act as "Alexa" (turning lights on or off, etc)
Cool bro🎉🎉🎉
Hi can i use MAX98357 I2S 3W CLASS D powerful AMPLIFIER instead of i2s amplifier?
Can we have programing test read out from this ? Of program lines
Can we use this model as the speaking model for my robot project ? Is it a good idea ....help me out plzz 😊
you are genius 👍👍
Thanks 😊
Bro in which platform ur writing code and in which language ur using plz tell me bro
Sir also make a video on google bard AI
How to make a voice controlling gps navigator saying directions as output in a speaker
Should we buy cloud for it
is it possible to use one microcontroller instead two microcontroller
hi, how to fix error I2S.cpp:57:10: error: 'i2s_read_bytes' was not declared in this scope
Is there a way to make this work with T-Display ESP32-S3 boards, would be cool if you could use both screens, the speech recognition board could show the text from your speech and the output board could show the text from GPT.
that is exactly what I was thinking
Can you remake the schematic diagram and code while adding a micro sd card and increasing the wavdatasize I'd really appreciate
where can i get more datail about this project like connections etc...
Can you make a video using Tensorflow lite?
hello,good job
Very good
Thanks
Can we control home appliances with it?
Bro. Bunun Türkçe dilinde nasıl yapılacağı konusunda yardımcı olurmusun?
how big of battery are u using????
can it be done just by one ESP32 module ?
Hi how can I make this project only using one board
Can you merge Google assistant and Chat GPT , ( yani Ok Google ya Hay Google bol kar Hame esa kahe Answer the Question using ChatGPT , to Google assistant hame ChatGPT redirect karde ya Usse Voice leke Read karne lage , overall ChatGPT+Google Assistant Also Murge Google Bard Ai with Google Assistant and make own Assistant
Teşekkürler.
Can you make it start by voice and not motion detection? Improve the delay time. Store common questions locally?
We tried to lower down the delay, but didn't get success in it as there are a lot of task running for getting the answer
Can i ask the question in malayalam, is it responde Or not?
How many dates to ship your product
Does it talk any language or English only thank you
Hi !! Greetings from Argentina !! No one had the abbility to solve my case, maybe you can. I bought two handys Baofeng BF999s, because they are cheap and transmit 500 mts. (it says 10 km......), 500 mts. is ok for me. But, I want to send data from one arduino to other arduino, via handy, that`s why I bought two. It only has 3.5 jack/plug and 2.5 jack/ plug, has no usb port, no numeric keypad. There is not clear information enough...... can you help me ??????? schematics and code, for arduino uno or mega . Thanks a lot !!!!!!!!!!!
hi bro i am doing this project for exibition . when it done copy your code insert the libraries it doesnt get chat gpt replay
have you provided the needed api? or you can use other models like (if OpenAI's api doesn't work) GeminiUltra by Google which is free (60 requests per min)
Add an lcd that displays the input and output speech
Can you add facial expressions using an OLED screen or any other screen? And also make it respond to a wake word like Hey GPT or any wake word
We are thinking of adding a wake up command but let's see
@@techiesms that’s great What about it’s personality
How can we modify this into a language translator between two different cultural people sir. I am working on a project and your answer will definitely helps me sir.
Hey I'm working on something almost similar, could we get in touch somehow?
Sir please make chatgpt and home automation
Does this project run on esp8266?
Can we select the language ? I have interested idea ! 😊
CHAT GPT already detect language, so no need to change.
🔥🔥🔥🔥
can't we rename the library?
Bro we are doing this project can u help us regarding this
while running the speech to text code I have got( 'LED_BUILTIN' was not declared in this scope ) in "pinMode(LED_BUILTIN,OUTPUT);" this line , how to rectify this?
I am stuck on this too. Any solution found?
Just select the board as ESP32 Dev Kit V1 in the boards section
@@techiesms Thanks, but had to comment out the led_builtin code to make it work further. But now I am getting this error - E (2623) i2s(legacy): i2s_check_set_mclk(1799): ESP32 only support to set GPIO0/GPIO1/GPIO3 as mclk signal, error GPIO number:3000 E (2624) i2s(legacy): i2s_set_pin(1881): mclk config failed Can you help with this? I will be happy to pay for your time. Thanks.
@@techiesms I was able to downgrade to v1.0.6 and the upload to ESP32 worked. However, I get the following error Press button Record start! Recording Completed. Now Processing... Json daata--- Guru Meditation Error: Core 1 panic'ed (LoadProhibited). Exception was unhandled.
There is an error in the code. The I2S.h library is faulty, the code from speech to text gives an error.
me too facing the same error...can any one help
Anyone who compiles and gets the error esp32 I2S error: 'i2s_read_bytes' was not declared in this scope, I suggest using ESP32 Board version 1.0.6
Please make again
I think the best idea is to use voice command to wake up the device instead of IR
Yea
Yes but its implementation is not as easy
Which component is there downside the battery charging module 5:35
HT 7333A
Can i use esp 8266
Woo
can we use a free API ??
Hi Sachin! upload the code for Text To Speech and Speech To Text to each ESP32?
or one of the esp32?
pinMode(LED_BUILTIN,OUTPUT); error while uploading speech to text in esp32 what to do ?
i am also getting the same error, how did you rectify this error ?
Hello, I too similiar problem
@@patrickrios99did you solved this problem?
Please detailed vedio applode and all components name or by link and full code link
What is cost?
Hlo brother if i orderd theese components. How many days it will came to my destination
What’s your order number ?
@@techiesms no I couldn't ordered but suppose if I order how many days it will come to my home
How could it be done with just one esp32?
Btw, amazing video, congrats!❤
Please add home automation chatgpt
nice glasses
Can we make a robot with this idea ?
Hi, sir I have completed hardware but all the code is having compiling error. Please help, as I have spent all my savings 😭
hello i m facing the similar problem, did u figure your one out? if yes pls help
@@goodsquirrel2725 bro did u find the errors solution Whats ur error i have error too
Urgent!! My project works fine, from the speech to text to chatGpt to text, but i don't get it out as an audio output, CAN anyone help. My amplifier and ny speakers are working fine. I dont know the issue, PLEASE HELP. THANKS
I am doing this project, but once I start everything I am getting guru meditation error in the serial monitor, can you please help me with this project
hi, i have problem to, do you solved?
@@fitraanantioramadhani8695 did you completed the project ?
Is it possible to prepare a robot to speak Arabic?
WARNING: library Audio claims to run on sam architecture(s) and may be incompatible with your current board which runs on esp32 architecture(s). \ChatGPT Voice Assistant ote: suggested alternative: 'Audio' audio.loop(); ^~~~~ Audio exit status 1 Compilation error: 'Audio' does not name a type showing error 🥲
did u make it
hello, did You get rid of the error u were facing. i am facing the same error. please help.