Free YouTube Transcribe

Video transcript

Build Your Own JARVIS with Jev (INSANE)

Henryk Brzozowski · 1,905 words · 9 min read

WhatsAppTelegramX

Want to search this transcript, jump the video from any line, or download it as TXT, SRT, or VTT?

Open in the transcript tool

Full transcript

Replacing Siri with Jev

0:00Hey Jeff, can you open up Spotify for

0:01me, please?

0:03>> There you go, Spotify.

0:07>> Hey Jeff, can you play the song for me

0:08in Spotify?

0:10>> Playing.

0:12>> Hey Jeff, can you turn down the volume

0:15in Spotify and then open up [music]

0:17Slack for me, please?

0:21>> All done.

0:23>> Amazing. Uh hey Jeff, can you set a

0:25reminder to call Bob in 5 seconds,

0:27[music] please?

0:30>> Got it. I'll remind you.

0:35Hey, it's time to give Bob a call.

0:38>> Hey Jeff, uh can you now uh tell me who

0:41wrote Hamlet?

0:42>> William Shakespeare wrote Hamlet, one of

0:44his most famous tragedies from the early

0:461600s.

0:48>> Uh hey Jeff, can you now uh

0:50close down or lock down my computer,

0:51please?

0:53That's cool.

0:54>> Screen's locked.

1:00>> And there you go, it's locked.

1:02How insane is that? Now, I'm just going

1:05to get it to close stuff. Hey Jeff, can

1:06you close down my Slack and Spotify?

1:09>> Both sorted.

1:11>> That's so It is so bloody cool. This is

1:13one of the coolest and best things I

1:15think I have ever built. In this video,

1:18I'm going to break down exactly on how

1:20you could use it and potentially on how

1:22you could build it yourself. So, stick

1:24around and let's get to it. All right,

1:27so to make this work, we need at the

The three pieces

1:28bare minimum these three main things.

1:32Number one is we need Fish Audio, which

1:34is the sponsor of this video, and the

1:36voice that you heard was Hannah, and

1:37she's got that like super chill,

1:39realistic voice. And if you guys would

1:41have want to check out some of the other

1:42ones, Celine sounds really cool, Sarah

1:44sounds uh awesome, too. You know, if you

1:46are thinking of using something like

1:47this, like a like a your your own

1:49Jarvis, and you want like a male voice,

1:51you could certainly come in here and

1:52choose a male voice. You could even

1:53create your own voice it within here.

1:56So, you could clone your voice, you just

1:57have to talk to it for like 1 minute. I

1:59actually recorded myself in here before

2:01and I cloned my voice. So, if I for

2:02example say Henrik, I'm going to

2:04generate speech.

2:06Henrik.

2:07Henrik. FYI, if you're thinking like,

2:09"Yeah, I'm going to need to pay for one

2:10of those subscriptions." No, that is not

2:12true. Right now, official audio is

2:13currently for free. So, you can use it

2:16at least up until the end of November

2:182026 using this specific line, S2.1 Pro

2:24free. I think the only thing you have to

2:25pay for is this next thing, which is

2:27Jev.

2:28>> My name is Yeah.

2:29>> And it is a way to help us grab our

2:32commands or questions into a spot where

2:35something creates decisions for us. And

2:38these decisions get then turned into

2:39action for us on our computer. So, for

2:41example, before I said, "Hey, can you

2:43open up Spotify for me, please?" And it

2:45opened that up. Then it created a

2:47decision for itself. Does the user want

2:49to open up Spotify? Answer is yes. Then

2:51it would go and open that up. But, "Hey,

2:53can you open up Spotify for me, please?"

2:55>> Spotify, opening now.

2:57>> And um if you guys do want to learn more

2:58about it, I'm sure there's like a crap

3:00ton of videos out there right now on

3:02explaining on what Jev is. It is

3:04essentially our back end that controls

3:07the actions that we want it to take.

3:09What I would recommend getting is this

The LLM fallback

3:12thing called Open Writer. From here,

3:13you're going to be able to get an API

3:15key. And this API key lets us set up a

3:17fallback for things like

3:20when I asked it for example about

3:22Hamlet, when it can't take actions on my

3:24computer and it's outside of like its

3:25scope. What it's going to do is think,

3:27"Hey, I can't really answer this

3:28question." So, then it's going to go and

3:31ask an actual LLM.

3:33So, you guys can choose any LLM that you

3:35like. You can choose Haiku, you can

3:36choose Gemini models, you can choose

3:38your llamas or whatever you want. And

3:40you can get the LLM to create you the

3:43response. Now, I also, if you notice for

3:45the

3:46for the timer, it actually responded

3:48back to me very nicely. So, for example,

3:50if I say, "Hey Jev, can you uh tell me

3:52to book an appointment in uh 3 seconds,

3:55please?"

3:56>> Sure. I'll give you a shout.

4:02Hey, it's time to book that appointment.

4:04>> particular case, that was a timer that

4:07got sent through

4:08uh Jeff, then that got sent through uh

4:11I'm actually using a Haiku 4.5 just cuz

4:13I really like that model. Now, to some

4:15of the use cases that I think would be

Use cases

4:17awesome for this. Um number one is what

4:19we were using just then, right? So,

4:22uh computer use / control.

4:25Uh later, I think this could be taken

4:27into like writing your own emails for

4:31you. I was just really excited,

4:33honestly, to get it to turn into an icon

4:36on my on my Mac.

4:38Uh have it open up in my dock. Um so,

4:40I'm sure you guys can think of uh more

4:43exciting things that you want to use it

4:44for your own computer use, for whatever

4:47you guys want to use that for. Another

4:48cool one would be like uh house control

4:51system. Instead of Siri or Alexa, you

4:53could be using this setup, which would

4:56be for free and running on your local

4:59setup, by the way. So,

5:01uh where when I'm running it from right

Local transcripts

5:04here,

5:05it's actually not going the or the

5:07transcript is staying on my machine. So,

5:10we're actually using I'm just going to

5:11write here.

5:12We're using a small Whisper model, which

5:15only takes up maybe like 250 MB to run.

5:19Do an uh check every so often to see

5:21whether I've said, "Okay, Jeff." or

5:23"Hey, Jeff." And by the way, that's

5:24totally uh changeable. So, anyway, uh

5:26going back, house control system, uh you

5:28know, you could say instead of me like

5:30saying, "Hey, hey, Jeff, open up

5:31Spotify." you could say, "Hey, Jeff, can

5:33you turn on the living room lights?"

5:34Then, you could potentially use it, you

5:37know, as like a uh disability kind of

5:39helper to help you control your computer

5:41a little bit better. However, for me,

5:43what I actually would like to use a

5:45four, be to what do multitask. Hey Jeff,

5:48can you open up Spotify for me while I'm

5:50smoking? I don't have to like come to my

5:52computer and you don't press buttons. So

5:54multitasking, opening up stuff, and

5:56potentially lessen the screen time.

5:59If you guys can think of any other cool

6:00use cases, just put them down in the

6:01comments below. Now, we are going to get

6:03into the juicy stuff. So, if you are a

6:07person loves building this stuff, loves

Build it yourself

6:09Claude code, Codex, stick around and I'm

6:11going to show you on more or less on how

6:14to put this together. So,

6:16number one step is you need an IDE. So,

6:20you could use Cursor, you could use any

6:22other IDE that you like. What I'm going

6:24to do is go down into the description

6:28and then grab this.

6:31Then,

6:32come back into Cursor, you're going to

6:34go clone repo, and then you're going to

6:36put that into any folder that you like.

6:38So, for example, I could put in my life

6:40get folder and I could select that as my

6:42repository destination. It's going to

6:43open that up for me.

6:45And now I have that project that I

6:47created for you guys. You can either

6:49follow the steps or you could just get

6:51your Claude code or Codex to set it up

6:53for you. So, for instance, you know, you

6:55would come into here, you would just

6:57type in Claude,

6:59and say yes.

7:01Then, you would grab this. You just

7:04copy this paragraph right here.

7:06You paste that in, press enter, and it's

7:08going to do all the magic for you. So,

7:11if you came here and you're like, "Hey

7:12Henrik, I'm using your Windows. Why the

7:14hell would I be watching you build it on

7:16your Mac?" Well,

7:18easy fix is,

7:20you know, you copy this

7:21and you say,

7:23uh

7:24"This guy was using uh this on his Mac

7:27and he everything so that it works here

7:28on my Windows desktop." And then just

7:29let it just let it rip. The things that

7:31you will need is your API key. So, I'm

7:33just going to show you real quick where

API keys

7:34they roughly exist. Uh you will then

7:36come into uh developer in Fish Audio.

7:39You will click on create API key. You

7:41click on create, copy that.

7:43Then go to typesafe.com, you apply if

7:46you don't have the if you don't have

7:48your API key yet. They exist right here

7:51though if you do. So you click on create

7:53API key, you create it, you copy that as

7:56well.

7:57Then you would come into open router,

7:59you'd come into something that looks

8:00like this. So you click on API keys on

8:02the left top right here. You would copy

8:05the API key. So once you got those three

8:08things, once you got your app working

8:10and you've you know you've discussed a

8:12little bit with the Codex and Claude,

8:13you come into this button right here

8:15that says keys. You press that and it

8:17opens up this new window

8:19you just paste all your respective keys

8:21into here. You press save. That gets

8:24saved to your Mac OS keychain. So like

8:26no LLM or whatever sees it so everything

8:28stays safe on your in your environment.

8:31So at the moment I believe Type Safe is

8:32actually for free. I haven't set up any

8:35payment methods on it. So I've been

8:38using a so I've used 1. 3 million tokens

8:41and it's cost me what's that? 4 cents? 5

8:44cents? I was telling you guys S2 1.3 is

Permissions

8:47now free for developers. So if you just

8:49I jump into here. I guess the only thing

8:51that I should really worry about is that

8:53when you create this app it's going to

8:54ask you for some permissions. So you

8:56might have to go and like click allow to

8:57use the microphone. Allow to open up

9:00Spotify for you. And then once you've

9:02got it working

9:04mine for example was put into a DIST

9:08folder as an icon like this. And then I

9:10just click on this and it just opens up

9:12the app for me. And then you can choose

9:14between hold option and then hey Jeff

9:16option. So thank you guys for watching.

9:18If you would like to sign up to Fish

Last thoughts

9:19Audio I will

9:21I will give you absolutely nothing for

9:22it. Is there's a referral link down

9:24below. If you're like someone out there

9:26who just doesn't want to build this

9:27stuff for yourself then want someone

9:29like me to come in and just help you do

9:30it. You're more than welcome to also

9:31just book a call with me as also

9:33available right below.

9:36Anyway, thanks guys for watching and see

9:37you next one. Peace.

Recently added transcripts

Browse the whole transcript library Browse all English transcripts

This transcript was generated from the captions YouTube publishes for this video. Get the transcript of any YouTube video atfreeyoutubetranscribe.com, free, unlimited, no sign-up.