00:00What type of gear thing is?
00:2612.
00:4113.
00:4918.
00:5118.
00:53So, me and my team, when we work on something, we don't have track of food or anything.
01:03So we usually go all in, we'll be working for a whole day and we were like, one day,
01:07we were like, shit, we didn't even have food one set today.
01:10And then we were like, what if we have a device that can know when we are hungry and automatically
01:14order.
01:15So, this felt like a lame idea, but then it was very interesting as well.
01:19So, we then figured out we can actually build it by using a stethoscope, we can detect
01:23the voices inside our stomach or in our body, right?
01:27So we used it and we just did a small prototype, it worked.
01:31So then we built a small 3D printed case and wrote a code which connected through, so Zomato
01:37also launched their MCP server, which means you can activate, access their platform directly
01:43through using an LLM model.
01:45So it was an opportunity at the same time, so we used it and it worked.
01:50So it was cool, so we made a fun video out of it and it went viral, which is not expected.
01:53We then worked very well.
01:54Now I'll just...
01:56It's good.
01:57It's good.
01:58Yeah.
01:58It was great.
02:00It's good.
02:00It's good.
02:01It was good.
02:02It was good.
02:03It was good.
02:05It was good.
02:05It was good.
02:06Oh.
02:07Ok.
02:08This stethoscope can record voices inside your heart sounds,
02:23or whatever is there, it has a diaphragm that basically access the sounds inside which a
02:29normal microphone cannot.
02:30For this we attach the small microphone and that microphone goes to a controller called
02:34the ESP32.
02:36So this controller will basically detect these sounds as an analog input and then when the
02:42sound is detected it will access a model, a python script in the local server.
02:48So ESP32 connects to Wi-Fi, it has Wi-Fi access, connects to the internet to the local server
02:53where we run this code which will write a prompt automatically in Claude.
02:57Claude is an LLM model, it will write a prompt saying that order us this, this, this food,
03:02access our Zomato MCP server, using the MCP server, access my Zomato app, see what orders
03:09I have preached before, what kind of orders food I like to have and based on that if my
03:15growl is a smaller sound then order a basic food, if my growl is a bigger sound then order
03:20more food that means I am more hungry.
03:22So it will see all those preferences, see the ratings of the restaurant, see my budget
03:26and after comparing all those things it will find the perfect dish and order it for me.
03:31So it's checking my order history, it's my order that has been.
03:36So I don't look at any scalable thing out of this, it's a fun project, it can be used, but
03:43I don't think so someone would use it everyday.
03:50I don't think so someone would use it everyday.
03:57I don't think.
03:58I don't think.
03:59I don't think.
04:00I don't think.
04:01I don't think.
04:02I don't think.
04:03I don't think.
04:04I don't think.
04:05I don't think.
04:06I don't think.
04:07I don't think.
04:08I don't think so.
04:09I don't think so.
04:10I don't think so.
04:11Anyone who wants to build something or anyone who feels like they want to learn but they're
04:14like, OK.
04:15anyone who wants to build something or anyone who feels like they want to learn but they're like
04:39okay I need this money for it or this idea needs to get this this much of money only then I'll build
04:46it that will come to you eventually money will come to you eventually or the output of that
04:51will come to you eventually you just keep on executing even if it's the smallest idea that
04:56comes to you just go on and just execute it
Comments