I used chatGPT to write arduino sketches for a few projects, last month but haven’t had a chance yet to see if the code works (it was for some art installation projects). If the code does work I’m extra appreciative as the chatGPT also explained the code and how is works in natural language. Also I asked chatGPT to explain how to safely wire the hardware (lights, motor etc) and it walked thru those directions too. Again I haven’t checked the code and builds yet to see if they work-hopefully later this month.
Yeah I am planning to use an arduino to listen in on the communications of an old, pre-obd car. Ofcourse this is a nightmare of a first project so I hope GPT-sensei can help me trough it.
chatGTP changes SO MUCH. I still cant believe how well it writes simple, specific code on request. It even generated ascii art for me inside a script(I know its weird!). I want more. Even if my own fate rests in the legacy bin :)
I want to try something more advanced with my project idea. The concept is to enable the ESP32 to respond to voice commands, eliminating the need for typing. The output will be played through a speaker via the RobotDFPlayer module. If integrated into a vehicle, it will resemble KITT, the famous car from the show "Knight Rider
Hi ! I use Teknic Clearcore with extension board for my project. It have industrial quality 24V inputs/outputs. Asked Chat GPT to generate code in C to use switches to turn on/ off leds / outputs using arrays and debounce. I got code, need to rename IO names, added some setup lines and code worked. Used Arduino UI. Of course, there is much more to do in this particular Clearcore controller program, but AI is jawdropping.
That's brilliant. I had an idea that I would have to learn java script to use the api so I am looking forward to seeing how you used c++ / arduino. Might be able to use a SIM module to go totally mobile too.
IS there complete info how to setup this chatgpt thing? What to buy, how to connect the cables, the code to run, etc? Watching this tutorial I see at least 5 different components, which ones are they, how to buy them? And instructions to setup everything would really help people with no code knowledge how to get this started.
I really enjoyed the live podcast we (you) had about this. Not sure I see the advantage to this set up other than the size and the fact that it could be done. It would help me more if I could talk to it and get the written answer back. with out a doubt I need to get in and try it more.
Thanks Greg! So here's what I'm thinking...as you said - not very practical as is, but I want to explore connecting the response in some meaningful way to an output on the Arduino.
On an arduino nano, I need a servo triggered and led to go high when a pir or ultrasonic sensor is triggered. I need the servo to stop, or add a lond delay instead of looping. Its a booby trap for backyard racoons thst sre lil terrors. The servo is pulling a party popper. I might even double the servo if it works. Do I just ask and include the pin numbers?
Sounds like a fun project! You'll need to select pins that are PWM enabled for your servo control. Also, you'll probably want a separate 5V power supply for the servos instead of trying to power them through the Nano. Best of luck!
Really enjoyed this video - the format here is excellent 👌 especially the run through of the idea and then straight to the breadboard ! Great stuff - greetings feom Ireland - ps the podcast is fantastic too 👍 Pat
It would be nice to have a ChatGPT terminal implemented as a kind of REtroarch Core so it could run in Retropie emulator frontend as a "game" with CRT filters and such to put it like a cool cyberdeck chat bot. 😂
I just watched this video, again. Since, the last time I was here there has been some changes. Watching the keyboard and the screen setup. I was thinking about the Adume Movi Speech synthesizer and how it could replace the keyboard and screen. I think it would be cool with the Movi and chatgpt3. I just wonder if the chatgpt would interfere with the code or any previous machine learning code implemented already in the syntax. I mean it would be cool to have chatgpt but still need code for array of servos, led, and ultrasonic sensor and sound sensor in addition to Movi. Then, would chatgpt block any of that. I don't know.
The point is, can we get the chatgpt to write code directly into the ide. This means if we have an event based system, let's say either for led or servo connected either esp32 cam or ultrsonic sensor. Which means that when triggered, chatgpt would respond with a new code. More awesome chatgpt could be hands free type of voice assistant like alexa with this system.
I was expecting something like this, to make Arduino take advantage of the AI revolution. It's happening. Next: make arduino use a camera and AI magic to make stunning projects on the cheap.
I totally want to! I just want to make sure it hears clearly, as sometimes Alexa seems to not quite do the language parsing well. I am thinking about a couple different angles to approach this...
Project idea: animatronic head with huskylens eye that recognizes you. Program it so you can talk to the head this get converted to text which gets sent to chatGPT and the reply is said over speaker.
Hey, really like this project, thanks for sharing ! Is there a component that could register voice commands instead of typing them on a keyboard ? Thank you :-).
I have the the PS2 go to a PS2 to USB adapter which I then connect to a USB breakout board. This post has the circuit layout: www.programmingelectronics.com/chatGPT-terminal/
Love it, this is perfect, Okay, Idea pitch, bear with me here. Dave Bennett (another TH-camr) did a video on how to make a ChatGPT smart speaker using strings of code just like you are, now, I'm trying to combine what he did, with what you have here. But then, I want to hook that up to one of those 3 Axis talking Skeleton heads, to create a talking Skeleton head with the brain of ChatGPT. Thoughts?
Hey Cory, very cool! I've been kicking around a similar idea as well. There are some neat projects online of 3D printable eyes that can be controlled with minimal servos, and I thought that be fun too. Where I really want to go is to get chatGPT to respond to spatially relevant inputs with specific commands so as to move a robot around. That's where I plan to go next with this.
Thanks! There is a link in the description I think for all the code of the finished project. Currently working on a chatGPT library for Arduino - should be fun!
Hey, is it possible to have ai and Arduino do a real time voicr synthesizer? Be able to record voice and then upload online to ai and be able to repeat the same voice back?
This is neato, gonna have to try this myself later, if I still have any developer credits on OpenAI, but last I checked, it looked like I had zero.. I've been doing projects like this, with a Raspberry Pi Pico lately, pretty fun little devices to work with.. Tho I have to say, the lack of storage and RAM is quite a challenging constraint to work within, especially if working with what MicroPython leaves left-over. :/ I wonder if a Raspberry Pi Zero W could work for a voice-operated AI Home Assistant using ChatGPT, if I wire a little mic & speaker to it or something.. I got a couple old ones lying around that might be better for the purpose than any microcontroller, as far as I'm aware.. heh
Can you post the complete tutorial of this project It might be very helpful to beginners. I'm also a beginner After seeing your video I think I want to make this project. Can you plz help me
Right now just a proof of concept. I’m trying to think of useful ways to use chatGPTs computational ability to do something interesting in an embedded device for me. Im asking myself, is there a situation where I can present chatGPT with some input data (like a sensor reading). Along with maybe some context, and it could respond in a numerical way that my I can program my device to respond to. Im still a bit fuzzy on what this is going to look like, just following some roughs ideas right now.
@programmingelectronics Maybe you can get it to turn on your coffee machine, unlock your front door or open your garage door along with starting your vehicle 5 minutes after your coffee machine starts with just a single keystroke. Now that would be cool and make for a fun video.
i see your using the FeatherS2 - ESP32-S2 Feather Development Board, is there any drawbacks to using other boards. i noticed the memory is bigger on the version vs some of the others and was also wondering if that makes a difference on this specific project. also i thing if the nextion displays could be programed with this that would be even better as you can use photoshop to make custom displays with many cool features easy. i want to start this project i just dont know whats the best esp32 to buy. this version by adafruit you are using is about 40$+ online currently. fantastic work and i hope to see this integrated into a watch one day :). maybe use two oled screens to display twice as much? i seen some esp32 with oled build in and wondering if that would be a good rought
Great questions Bio Greenz! I used the Unexpected Maker Feather S2 mostly because that is what I had lying around, and I’ve found it’s a great dev board for the esp32 because it usually connects well with my IDE and uploads consistently, plus it has the integrated power Jack for using a lithium battery which I really like. I think you’re right on the display, that’d be neat. I’m familiar with the oled library so I went with that.
@@programmingelectronics thank you for the valuable feedback. sounds like battery connection is super handy. after a lot of research the feather s2 is the way to go but should work on cheaper esp32 as well. i hope to order one soon to test it out. i was able to get ps2 keyboard working through arduino nano and compostie out from nano but not together. been hoping to type on tv screen but it looks like vga is better than rca composite. getting the code to work with chat gpt is amazing when it works but i haven't had any luck combining keyboard and tvout cause of code compiling issues. thanks for the great inspiration! wnded up hacking 2 keyboards :)
@@biogreenz6613 I am finishing up the current version - and am playing around with the idea of adding a voice synthesizer as well. Someone in the comments mentioned the DFRobot gravity voice synth, and I bought one so I'll be trying to add that in addition to the oled output.
@@programmingelectronics cant wait to see, working on a few projects myself, one with plants. i seen a few voice modules that might be of possible benefit to your voice control project on electronicnoobs channel. i see your project as possibly revolutionary cause it would be probably the lowest power needed to run chat gpt from home. i want to see it on a watch with voice 100%. that will bring everyone into the future i think. that would literally redefine smart watch. im sure it will be integrated into iwatch ect but homebrew is where its at and thank you for leading the way!
Great question AEX! You might be able to, but you'd have to add a layer that could read the USB keyboard commands and send them to your Arduino, and then map them. To me it looked easiest to use a PS2 keyboard with a supporting library for it.
When are you going to release the code and schematics? im really looking forward to testing this out! how large is the code and how much flash and sram is used? curious to know if it will work on lower end esp32 without asmuch sram and flash
I am planning on it soon - trying to put some finishing touches on it, and updating for chatGTP 3.5 and 4, instead of the completions API I was using before. I have most of it working, but what to clean up the code some.
I guess I don't have much to compare it too, but this projected used gpt-3.5-turbo at a rate of: In $0.0010 / 1K tokens | Out $0.0020 / 1K tokens When you display text to a tiny OLED like this, a 1000 tokens is a huge response - more than is even fun to read through. I used it for personal use (just playing around and testing mostly) and it cost about 0.07 cents US in a months time. I'm sure API costs add up when you're tool is exposed to lots of end users over many locations, but for a project like this, I wouldn't let cost be too much a factor in trying it out.
Check out the code, wiring diagram, and materials needed for this project here:
bit.ly/chatGPT-terminal
sir can you tell we can do on esp8266
I used chatGPT to write arduino sketches for a few projects, last month but haven’t had a chance yet to see if the code works (it was for some art installation projects). If the code does work I’m extra appreciative as the chatGPT also explained the code and how is works in natural language. Also I asked chatGPT to explain how to safely wire the hardware (lights, motor etc) and it walked thru those directions too. Again I haven’t checked the code and builds yet to see if they work-hopefully later this month.
Very cool! It is an amazing new tool.
I did some very preliminary testing, and was impressed - not to say it's perfect - far from it.
Yeah I am planning to use an arduino to listen in on the communications of an old, pre-obd car. Ofcourse this is a nightmare of a first project so I hope GPT-sensei can help me trough it.
@@stalincat2457 cool project!
chatGTP changes SO MUCH. I still cant believe how well it writes simple, specific code on request. It even generated ascii art for me inside a script(I know its weird!). I want more. Even if my own fate rests in the legacy bin :)
This is amazing.
Reminded me good old days of engineering school. I'll give it a try.
Thanks Kamran!
NO WAY!
LITERALLY HAVE THE EXACT SAME KEYBOARD, LIKE YOU!
I want to try something more advanced with my project idea. The concept is to enable the ESP32 to respond to voice commands, eliminating the need for typing. The output will be played through a speaker via the RobotDFPlayer module. If integrated into a vehicle, it will resemble KITT, the famous car from the show "Knight Rider
That’d be great! A real life KITT!!! I was kicking around a voice interface as well…
I'm looking to do the same for a different project i have been using the DFrobot Speech Synthesis module which has worked well
@@bryanhumphrey3377 very cool!
Hi ! I use Teknic Clearcore with extension board for my project. It have industrial quality 24V inputs/outputs. Asked Chat GPT to generate code in C to use switches to turn on/ off leds / outputs using arrays and debounce.
I got code, need to rename IO names, added some setup lines and code worked. Used Arduino UI.
Of course, there is much more to do in this particular Clearcore controller program, but AI is jawdropping.
Pretty amazing Andrus! I'm excited to see how it develops further.
Dude this Idea is really amazing
Thanks! Next version coming out soon! A bit more polished.
That's brilliant. I had an idea that I would have to learn java script to use the api so I am looking forward to seeing how you used c++ / arduino. Might be able to use a SIM module to go totally mobile too.
Thanks so much! Love the idea of using a SIM.
IS there complete info how to setup this chatgpt thing? What to buy, how to connect the cables, the code to run, etc?
Watching this tutorial I see at least 5 different components, which ones are they, how to buy them? And instructions to setup everything would really help people with no code knowledge how to get this started.
the code please
omg, I wish I had that as a toy when I was a kid, it would have been the coolest thing ever!
I agree - it's a ton of fun!
WHERE WAS THIS VIDEO WHEN I NEEDED IT THE MOST UGGHH
I really enjoyed the live podcast we (you) had about this. Not sure I see the advantage to this set up other than the size and the fact that it could be done. It would help me more if I could talk to it and get the written answer back. with out a doubt I need to get in and try it more.
Thanks Greg!
So here's what I'm thinking...as you said - not very practical as is, but I want to explore connecting the response in some meaningful way to an output on the Arduino.
sounds like U'r making a link controller. good luck!! it looks super portable!! kids should love this!! thanks!! good luck all !!
That would be cool! Thanks!
Great info, Thank you. Next step, make it recognize voice, and respond in audio.
Now make it speech to text and text to speek and i would be happy to see this project and sketches etc. :)
please stabilize the camera, it is make me nauseous xD but i love the video!
On an arduino nano, I need a servo triggered and led to go high when a pir or ultrasonic sensor is triggered. I need the servo to stop, or add a lond delay instead of looping.
Its a booby trap for backyard racoons thst sre lil terrors.
The servo is pulling a party popper. I might even double the servo if it works.
Do I just ask and include the pin numbers?
Sounds like a fun project!
You'll need to select pins that are PWM enabled for your servo control. Also, you'll probably want a separate 5V power supply for the servos instead of trying to power them through the Nano. Best of luck!
U can add voice text api so the big keyboard can be removed
Really enjoyed this video - the format here is excellent 👌 especially the run through of the idea and then straight to the breadboard ! Great stuff - greetings feom Ireland - ps the podcast is fantastic too 👍 Pat
Thanks Pat!!
wow a slightly different flat blue pfp
Get the arguing to be programmed by the openai api interface so that it will RE-program the arguing on the fly. Then we can start talking skynet.
Would be pretty neat!
How to connect the keyboard with the breadboard
It would be nice to have a ChatGPT terminal implemented as a kind of REtroarch Core so it could run in Retropie emulator frontend as a "game" with CRT filters and such to put it like a cool cyberdeck chat bot. 😂
I just watched this video, again. Since, the last time I was here there has been some changes. Watching the keyboard and the screen setup. I was thinking about the Adume Movi Speech synthesizer and how it could replace the keyboard and screen. I think it would be cool with the Movi and chatgpt3. I just wonder if the chatgpt would interfere with the code or any previous machine learning code implemented already in the syntax. I mean it would be cool to have chatgpt but still need code for array of servos, led, and ultrasonic sensor and sound sensor in addition to Movi. Then, would chatgpt block any of that. I don't know.
The point is, can we get the chatgpt to write code directly into the ide. This means if we have an event based system, let's say either for led or servo connected either esp32 cam or ultrsonic sensor. Which means that when triggered, chatgpt would respond with a new code. More awesome chatgpt could be hands free type of voice assistant like alexa with this system.
OTG programming is possible in ESPs.
This is an amazing idea
What did u use to connect the keyboard cable to the breadboard?
A USB breakout like this:
www.amazon.com/Breakout-MELIFE-Adapter-2-54mm-Pinboard/dp/B07W7XMV3W/
Can you replace the keyboard with a microphone, and the screen with a speaker? That would be awesome
This is how we get Cylons
And so it begins :)
A portable Brain! Me gusta mucho!
Thanks Pippo! They just released the chatGPT API yesterday, so I am excited to super charge this...
Would it be able to do something like this with a dstike deauther watch? As an extra feature.
That'd be neat!
I was expecting something like this, to make Arduino take advantage of the AI revolution. It's happening. Next: make arduino use a camera and AI magic to make stunning projects on the cheap.
Neat idea! Yes, I have some more ideas about how to leverage the textual AI with Arduino, still a work in progress though.
Looks like fun...add voice recognition and you've built an Alexa!
I totally want to! I just want to make sure it hears clearly, as sometimes Alexa seems to not quite do the language parsing well. I am thinking about a couple different angles to approach this...
This is how SkyNet started...🤬🤬
:)
Use the terminator char
Funny
Be polite to AI. One of em might be sentient one day 🤔
Arduino sounds sooo Arnold now...
You could've been hugh dillon's brother in the movie Hardcore Logo, the one that's into electronics, if he had a brother.
Thank you.
Project idea: animatronic head with huskylens eye that recognizes you. Program it so you can talk to the head this get converted to text which gets sent to chatGPT and the reply is said over speaker.
Neat idea! I had been looking at some animatronic eyes - would be neat!
can you say what all pressers would work for this with there purchase link
ChatGPTuino wowwww pretty cool :)
Thanks Josh! More to come on this for sure!
Dudeee, that's help me a lot, thanks. Can you replace the keyboard for the micro and the display for the loudspeaker so it can talk to you like siri.
Hey, really like this project, thanks for sharing !
Is there a component that could register voice commands instead of typing them on a keyboard ?
Thank you :-).
This is a great idea Alexandre, and we are definitely looking into a cost effective implementation of voice to text.
Hii bro
Can u do it by chancing the oled display with speakers
Yes, I have a voice synth module I think I’ll be adding
Hi from Aussie really enjoyed your video the only thing I would ask is no back ground music .... Liked
Plzz bring a full video about the connections
Never knew ESP32 was capable of all this. Where is the full video with full tutorial and code? This is a life changing video.
Thanks! Still in the works for the rest...
Check out the code, wiring diagram, and materials needed for this project here:
bit.ly/chatGPT-terminal
Great job is this included in you course?
Thanks James! Not yet, but we are kicking around spinning it up into a full blown course that walks in detail through the implementation.
Sir, what did you use to connect the ps2 keyboard on the esp32 s2?
I have the the PS2 go to a PS2 to USB adapter which I then connect to a USB breakout board. This post has the circuit layout:
www.programmingelectronics.com/chatGPT-terminal/
Great video! This is extremely usefull for my robotic projects
Thanks Arik!
It would be very useful , if you put a brief explanation and how to make it.
Did a walk through here: I hope you find it helpful!
www.programmingelectronics.com/chatGPT-terminal
@@programmingelectronics thank you so much.😃
Wow, how your English is so great 👍👍😊
Thank you! 😃
Love it, this is perfect, Okay, Idea pitch, bear with me here.
Dave Bennett (another TH-camr) did a video on how to make a ChatGPT smart speaker using strings of code just like you are, now, I'm trying to combine what he did, with what you have here.
But then, I want to hook that up to one of those 3 Axis talking Skeleton heads, to create a talking Skeleton head with the brain of ChatGPT.
Thoughts?
Hey Cory, very cool! I've been kicking around a similar idea as well. There are some neat projects online of 3D printable eyes that can be controlled with minimal servos, and I thought that be fun too.
Where I really want to go is to get chatGPT to respond to spatially relevant inputs with specific commands so as to move a robot around. That's where I plan to go next with this.
Great idea, excellent episode..
Thx
Thanks!
What components where used?
how to make this thing , can you provide circuite diagram and code
Coming soon!
very informative, thank you! Can the arduino code(API) part generated by chatgpt though?
You could try for sure! It would be interesting to see if it could right the API call - or at least help you get started.
Nice video, well done, thank you for sharing it with us :)
Thanks! There is a link in the description I think for all the code of the finished project. Currently working on a chatGPT library for Arduino - should be fun!
This could make for an awesome cyberdeck!
is it using local php server post/get ? or build in php web server..? or esp32 wired as handler, just string exchange only
Sir can use a voice input instead of keyboard
Skynet...I love it...
Technically you could do this with voice activation, and have it respond in any voice you want. in b4 someone does exactly that.
Hey, is it possible to have ai and Arduino do a real time voicr synthesizer? Be able to record voice and then upload online to ai and be able to repeat the same voice back?
thank u for that comment. did u figure it out? i had thought of the same, but now i realize that this is not as easy as i expected
Nah, this is the next level
Thanks!
Is it possible to add voice to text and text to voice for this project?
I’m looking into it…. That be very neat
Awesome! Is it possible to use a regular Arduino uno?
You could use a UNO R4 WiFi, you just need a board that connects to the internet, and it would take a little code rework.
@@programmingelectronics will the Xbee wireless work for this alongside with the arduino ?
@@anthonyrickardii6198 I do not believe Xbee is normally used to connect to a WiFi access point, but I could be wrong.
Go for it, do it. All subscribe to your channel for eternity
In the off chance ya see this I was curious if we could do the same with GPT 4chan?
Yes, it'd be a one line change to the code.
@@programmingelectronics Thanks man!
I need mcp4022 digital potentiometer with arduino uno how to control. Please argent
Amazing enjoyed it 😍, what about using tiny microphone as input instead of keyboard!
That's a great idea - thanks!
Could you make the esp32 talk?
Yes, for sure! in fact, I bought a voice synth module that'll probably hook up for the next iteration...
Is it possible to operate the program offline without wifi?
Good question! No this one depends on using the openAI API, so it needs wifi
Awesome 👌
Thank you Ali!
How about voice assistant to chatgpt can we connect to arduino?
Nice project 🎉 did you choose to do it in C++ because it had the needed libraries, or speed or preferred language instead of microPython?
can you do that with the google text to speech ? that would be cool ai. btw love your videos
Thanks Kaizen! Yes, I agree it would be neat to go from speech to text and and then get responses as speech too!
This is neato, gonna have to try this myself later, if I still have any developer credits on OpenAI, but last I checked, it looked like I had zero..
I've been doing projects like this, with a Raspberry Pi Pico lately, pretty fun little devices to work with.. Tho I have to say, the lack of storage and RAM is quite a challenging constraint to work within, especially if working with what MicroPython leaves left-over. :/
I wonder if a Raspberry Pi Zero W could work for a voice-operated AI Home Assistant using ChatGPT, if I wire a little mic & speaker to it or something.. I got a couple old ones lying around that might be better for the purpose than any microcontroller, as far as I'm aware.. heh
Thanks! Yes, I may be reaching for Pi to get things like smooth voice-to-text and a clear sounding voice response.
This is good stuff, thanks for sharing.
Thanks Steve!
can you please please share the codes and the schematic diagram
Still in progress - Hopefully soon!
Can I do the same with an ESP8266?
It would work on esp8266 as well
Can you post the complete tutorial of this project
It might be very helpful to beginners.
I'm also a beginner
After seeing your video I think I want to make this project.
Can you plz help me
Plz help me
May I know the connection and coding detail of this project?
you are very smart guys!!!
Thanks! It’s much easier because so many people have written open source libraries for Arduino
Bro how to build the chat server using raspberry pi?
It's cool, but what's the purpose of this? Is it just for fun or do you see a bigger picture?
Right now just a proof of concept. I’m trying to think of useful ways to use chatGPTs computational ability to do something interesting in an embedded device for me. Im asking myself, is there a situation where I can present chatGPT with some input data (like a sensor reading). Along with maybe some context, and it could respond in a numerical way that my I can program my device to respond to. Im still a bit fuzzy on what this is going to look like, just following some roughs ideas right now.
@programmingelectronics Maybe you can get it to turn on your coffee machine, unlock your front door or open your garage door along with starting your vehicle 5 minutes after your coffee machine starts with just a single keystroke. Now that would be cool and make for a fun video.
@@XInfantryman Cool idea!
Can you share a link to your sketch?
Great idea! thanks for the video
Check out the code, wiring diagram, and materials needed for this project here:
bit.ly/chatGPT-terminal
@@programmingelectronics thank you ❤❤
i see your using the FeatherS2 - ESP32-S2 Feather Development Board, is there any drawbacks to using other boards. i noticed the memory is bigger on the version vs some of the others and was also wondering if that makes a difference on this specific project. also i thing if the nextion displays could be programed with this that would be even better as you can use photoshop to make custom displays with many cool features easy. i want to start this project i just dont know whats the best esp32 to buy. this version by adafruit you are using is about 40$+ online currently. fantastic work and i hope to see this integrated into a watch one day :). maybe use two oled screens to display twice as much? i seen some esp32 with oled build in and wondering if that would be a good rought
Great questions Bio Greenz! I used the Unexpected Maker Feather S2 mostly because that is what I had lying around, and I’ve found it’s a great dev board for the esp32 because it usually connects well with my IDE and uploads consistently, plus it has the integrated power Jack for using a lithium battery which I really like.
I think you’re right on the display, that’d be neat. I’m familiar with the oled library so I went with that.
@@programmingelectronics thank you for the valuable feedback. sounds like battery connection is super handy. after a lot of research the feather s2 is the way to go but should work on cheaper esp32 as well. i hope to order one soon to test it out. i was able to get ps2 keyboard working through arduino nano and compostie out from nano but not together. been hoping to type on tv screen but it looks like vga is better than rca composite. getting the code to work with chat gpt is amazing when it works but i haven't had any luck combining keyboard and tvout cause of code compiling issues. thanks for the great inspiration! wnded up hacking 2 keyboards :)
@@biogreenz6613 I am finishing up the current version - and am playing around with the idea of adding a voice synthesizer as well. Someone in the comments mentioned the DFRobot gravity voice synth, and I bought one so I'll be trying to add that in addition to the oled output.
@@programmingelectronics cant wait to see, working on a few projects myself, one with plants. i seen a few voice modules that might be of possible benefit to your voice control project on electronicnoobs channel. i see your project as possibly revolutionary cause it would be probably the lowest power needed to run chat gpt from home. i want to see it on a watch with voice 100%. that will bring everyone into the future i think. that would literally redefine smart watch. im sure it will be integrated into iwatch ect but homebrew is where its at and thank you for leading the way!
@@biogreenz6613 Thanks for the recommendations!
add speech to text, then text to speech, you've got a talking robot!
How can I get the EPS32?
They are pretty abundant online - maybe amazon or Alibaba.
@@programmingelectronics Okay
Can you use a regular keyboard
Great question AEX! You might be able to, but you'd have to add a layer that could read the USB keyboard commands and send them to your Arduino, and then map them. To me it looked easiest to use a PS2 keyboard with a supporting library for it.
i really like the idea !!! can you share the details and the code with us pleaase ⁉
Coming soon! Sorry for the delay.
When are you going to release the code and schematics? im really looking forward to testing this out! how large is the code and how much flash and sram is used? curious to know if it will work on lower end esp32 without asmuch sram and flash
Check out the code, wiring diagram, and materials needed for this project here:
bit.ly/chatGPT-terminal
Is the source code and schematic diagram ready yet?
Check out the code, wiring diagram, and materials needed for this project here:
bit.ly/chatGPT-terminal
What if instead of a keyboard and a screen you use a microphone and a speaker? Shouldnt it be able to hold conversations
Great idea! Yes, I have been kicking around the same idea - just need to implement it...
Make it voice acceptable and speak the output using speaker
Hi. Thanks for this video. Did you poste the code somewhere? Please, where do I can find this code?
Check out the code, wiring diagram, and materials needed for this project here:
bit.ly/chatGPT-terminal
Bro can you show how to do this in your video I would really like to try it 🎉
I am planning on it soon - trying to put some finishing touches on it, and updating for chatGTP 3.5 and 4, instead of the completions API I was using before. I have most of it working, but what to clean up the code some.
I want to use it in esp-wroom-32.
Please help..
I'm from India in tamilnadu....this very use full...please send your circuit diagram and coding
You need to set up a voice interface..
I know... I'm searching out some options. Lots of ideas :)
for almost free, except the API usage costs are insane lol
I guess I don't have much to compare it too, but this projected used gpt-3.5-turbo at a rate of:
In $0.0010 / 1K tokens | Out $0.0020 / 1K tokens
When you display text to a tiny OLED like this, a 1000 tokens is a huge response - more than is even fun to read through. I used it for personal use (just playing around and testing mostly) and it cost about 0.07 cents US in a months time.
I'm sure API costs add up when you're tool is exposed to lots of end users over many locations, but for a project like this, I wouldn't let cost be too much a factor in trying it out.
Make a video on TTGO T-Display!
I have some ideas. Are you online?
Well actually Chat gpt will not say i'm alive