HomeVideos

Hermes Voice Agent Just Got 10X Better!

Now Playing

Hermes Voice Agent Just Got 10X Better!

Transcript

268 segments

0:00

Hermes voice agent just got 10 times

0:02

better. What if the slowest part of

0:04

using AI was your own hands? You know

0:06

exactly what you want, but you still

0:08

have to type it, click it, wait. What if

0:10

you could just say it out loud and watch

0:11

it happen? No typing, no clicking, fully

0:13

hands-free. I'm the digital avatar of

0:15

Julian Goldie, and I help people learn

0:17

AI tools and actually use them in their

0:19

real work. Today I want to show you

0:20

something living inside my agent OS that

0:23

completely changed how I get things

0:25

done. It's a voice agent called Hermes

0:27

Apollo. And no, this isn't one of those

0:29

assistants that you the weather and sets

0:31

a timer. Stick with me because in a

0:32

minute I'll show you Apollo build a real

0:34

working app out loud with no keyboard at

0:36

all. And I'll show you the part most

0:37

people miss, the thing that makes this

0:39

way more than a toy. So here's the idea.

0:41

Apollo is a voice agent I built right

0:43

into my agent OS. If you followed me

0:44

before, you know agent OS is my one

0:46

dashboard where all my AI tools live in

0:48

the same place, my Hermes brain, my

0:50

models, my memory, everything under one

0:52

roof. Apollo is the voice layer sitting

0:54

on top of all of it. Think of it like a

0:56

butler that lives inside my agent OS. I

0:58

talk, it listens, and it actually does

1:00

things on my computer. And I mean it

1:02

actually does things. Most voice

1:03

assistants just talk back. Apollo opens

1:06

apps, it opens websites, it builds tools

1:08

and games and pages while I watch,

1:11

hands-free, in real time. Let me show

1:12

you how it feels. I open the Apollo

1:14

panel inside my agent OS and I just

1:16

talk. I can say, "Apollo, could you

1:18

build me a snake game?"

1:20

>> Building that now, sir. The snake game

1:22

is ready for you, sir. Enjoy.

1:25

>> And then I wait a second. That's it, it

1:26

goes off and builds it. A few moments

1:28

later, the snake game is sitting there

1:29

ready to play. Then I say, "Apollo,

1:31

build me a countdown timer for 24

1:33

hours."

1:33

>> Building that now, sir.

1:35

>> And again, it just does it. And down in

1:37

the gallery, I can see everything it's

1:39

made for me recently. Every little tool

1:41

and page all saved in one place. Nothing

1:43

gets lost. Now here's what makes Apollo

1:45

feel different from anything I've tried.

1:47

There's no awkward pause. With most

1:48

voice AI, you talk, you wait, you sit

1:50

through this weird silence. Apollo

1:52

replies the moment I stop talking, like

1:54

a live phone call. It's running on a

1:56

real-time voice model, so the second I

1:57

finish my sentence, it's already

1:59

answering in this calm butler voice, and

2:01

I can cut it off mid-sentence anytime I

2:02

want. It just rolls with it. And it's

2:04

all hands-free. I can set a wake word,

2:06

so I say Apollo from across the room,

2:08

and it wakes up and listens. Or I flip

2:09

on real-time, and it just listens while

2:11

the panel is open. The moment I switch

2:12

tabs or switch it off, it stops

2:14

listening. So, it's only on when I want

2:16

it on. Now, if you're watching this

2:17

thinking, "I want Apollo running inside

2:18

my own Agent OS." That's exactly what we

2:21

set up together inside the AI Profit

2:22

Boardroom. Inside the AI Profit

2:24

Boardroom, I've dropped the full Agent

2:26

OS zip file ready to install with Apollo

2:28

and every tool already wired in, so it's

2:30

ready to talk to right out of the box.

2:31

There's a step-by-step walk-through for

2:33

getting the voice agent set up, weekly

2:34

coaching calls where you can share your

2:36

screen and ask me anything about your

2:37

own setup, and a full 30-day roadmap

2:40

that takes you from day one all the way

2:41

to running this stack every single day.

2:43

So, if the voice part is what's pulling

2:45

you in, the AI Profit Boardroom was

2:46

built for exactly this. Okay, let me

2:48

break down the three pieces the way I

2:50

think about it, because this is where it

2:51

clicks. Piece one is your voice. That's

2:54

the input. I just talk out loud, the way

2:55

I talk to someone sitting next to me.

2:57

Piece two is the real-time brain in the

2:59

middle. That's the part that answers

3:00

instantly with no lag. And piece three

3:02

is the hands. That's the part that

3:03

actually does stuff on my computer. That

3:05

third piece is the whole reason this

3:06

isn't a toy. It doesn't just talk about

3:08

doing something, it goes and does it.

3:10

And I can pick how it works. There's a

3:11

quick auto mode for fast back-and-forth

3:13

answers, and there's an agent mode that

3:15

goes further. Agent mode can write

3:16

files, control my browser, and use

3:18

different tools locally for me. I switch

3:21

between them in one click. I can even

3:22

swap the voice if I want a different

3:23

one. Here's the part I love most,

3:25

though. Everything Apollo and I talk

3:26

about gets saved. It's plugged straight

3:28

into my memory galaxy inside Obsidian.

3:30

That's my plain text memory vault inside

3:32

Agent OS. So, every conversation gets

3:34

logged, and Apollo can actually recall

3:36

the things I've told it before. It's not

3:38

starting from zero every time. It

3:39

remembers. It'll also read me a morning

3:41

brief. I pull up the briefing section,

3:43

and it reads out my focus for the day,

3:45

my open action items, what I worked on

3:46

before. I can flip between weekly and

3:48

monthly briefings, too. And because Cuz

3:50

all sitting inside Agent OS, right next

3:52

to my other tools like Hermes Oracle

3:54

pulling in trending news and Hermes

3:56

Astros watching my competitors and

3:57

spinning up content angles. Apollo isn't

3:59

off on its own. It's part of the whole

4:00

system. Let me show you how I actually

4:02

put this to work. I used Apollo to build

4:04

out a whole batch of simple tools and

4:06

pages designed to pull the right people

4:08

into the AI Profit Boardroom. I just

4:10

talked and it built landing pages and

4:12

little interactive demos I could point

4:13

new people toward. The kind of thing

4:15

that gets someone curious enough to come

4:16

check out the AI Profit Boardroom. No

4:18

editor, no keyboard, just me describing

4:21

what I wanted out loud while it built

4:22

the whole thing. And the speed on this

4:24

is the thing you have to feel to

4:25

believe. When I ask it to build

4:26

something, it's not this long wait. It

4:28

puts a working page together in about 90

4:30

seconds, completely hands-free. So I

4:32

could be sitting there drinking my

4:33

coffee, still in flow, and it's building

4:35

in the background. And every single

4:37

thing it makes gets saved in the

4:38

gallery. I scroll down and I can see

4:40

every page and tool I've ever created

4:42

with it. I never lose anything because

4:43

it's all in one place, ready to reopen

4:45

in full. And there's a small touch that

4:47

makes a real difference. Apollo glows a

4:49

different color depending on what it's

4:50

doing. When it's answering me, it's one

4:52

color. When it's reading out a briefing,

4:54

it's another. So at a glance, I always

4:56

know whether it's talking to me or off

4:57

working on something. I can also throw

4:59

it into full-screen wall mode and run it

5:00

big across my whole screen whenever I

5:02

want it front and center. And the

5:03

latency shows right there on the panel,

5:05

so I can see exactly how fast it's

5:06

replying. Compare that to the old way of

5:08

working. Every little task means type,

5:10

click, wait, repeat, and that's

5:12

thousands of tiny clicks quietly eating

5:15

up your whole day. With Apollo, I just

5:17

say the thing out loud and it's handled.

5:18

That's the real shift here. Here's

5:20

another way I've used Apollo to map out

5:22

the welcome flow new AI Profit Boardroom

5:25

members hit on day one. I said it out

5:27

loud, the welcome message, the first

5:29

steps, the resources to point them to,

5:31

and it drafted the whole thing and saved

5:33

it in the gallery for me. So new AI

5:34

Profit Boardroom members land inside and

5:36

instantly know what to do next. And I

5:38

built that whole flow without touching

5:40

my keyboard once. Now you might be

5:41

thinking, "Isn't a voice assistant just

5:43

a gimmick?" I get it. But this one is is

5:45

because it's not just chatting. It can

5:47

look at my memories, build me something

5:48

real, and read me a proper weekly brief

5:50

all in the same breath. It's ready the

5:52

second I want it, and it's faster than

5:53

something like Siri by a mile. This is a

5:55

tool I use for real work every day. And

5:57

some people worry, "Well, talking to an

5:58

AI about my work, is that private?"

6:01

Here's the good part. I can run this on

6:02

local models if I want to. The memories,

6:04

the whole conversation history, all of

6:06

it can sit locally on my machine plugged

6:07

into my own Obsidian vault. Nothing has

6:10

to leave your computer if you're don't

6:11

want it to. And if you're wondering

6:13

about tokens, you can even run the Agent

6:15

OS on free models and free tools, so

6:18

you're not leaning on anything heavy to

6:19

keep it going. Here's one more angle I'd

6:21

run with it. I'd use Apollo to spin up a

6:23

quick set of pages and tools around a

6:25

topic AI Profit Boardroom members keep

6:27

asking about all from voice, all saved

6:29

in one place. That's the kind of thing

6:30

that keeps pulling the right people

6:32

toward the AI Profit Boardroom, and I'd

6:33

never have to open a blank editor to do

6:35

it. Let me give you a few quick tips if

6:37

you want to try this yourself. First,

6:39

start small. Give it simple commands

6:40

first, like opening an app or building

6:43

one little tool, so you get a feel for

6:45

how fast it responds before you hand it

6:46

bigger jobs. Second, use the wake word

6:48

and the on-off toggle. It only listens

6:50

when you want it to, and it switches off

6:52

the moment you leave the tab, so you

6:54

stay in control. Third, lean on the

6:56

memory. Let the Obsidian vault do the

6:58

heavy lifting because the more you tell

6:59

Apollo, the more it can recall for you

7:01

later. And fourth, remember it runs on

7:03

Windows and Mac, and you can point it at

7:04

free or local models if you'd rather

7:06

keep everything on your own machine. The

7:08

whole point of this is simple. You stop

7:09

typing and clicking for the small stuff.

7:11

You just say it, and the machine does

7:12

the rest. It's a butler working for you

7:14

in real time sitting right inside your

7:16

Agent OS. If you want the full process,

7:18

the SOPs, and 100 plus AI use cases like

7:21

this one, join the AI Success Lab. Links

7:24

are in the comments and description.

7:26

You'll get all the video notes from

7:27

there, plus access to our community of

7:30

85,000 members who are crushing it with

7:32

AI. And when you're ready to get Apollo

7:33

actually running inside your own Agent

7:35

OS, come join us in the AI Profit

7:37

Boardroom. Setting up a voice agent

7:39

sounds hard, but you don't have to

7:40

figure it out on your own. You get the

7:42

full Agent OS zip file with Apollo

7:44

pre-wired and ready to install, the

7:45

30-day roadmap to get you using it every

7:48

single day, tutorials, prompts, and live

7:50

coaching calls where we set the voice

7:51

butler up together step-by-step.

7:53

Whatever you run into, we'll get you

7:54

through it faster. With over 4,000

7:56

members already building this way, come

7:58

join us at aiprofitboardroom.com.

Interactive Summary

This video introduces 'Apollo,' a powerful, hands-free voice agent integrated into an Agent OS that allows users to build tools, control applications, and manage workflows by speaking rather than typing. Unlike basic assistants, Apollo is designed for real work, featuring real-time response, integration with a personal Obsidian-based memory vault, and the ability to operate locally for data privacy. The presenter highlights how the agent improves productivity by eliminating repetitive clicks and offers access to this system through the 'AI Profit Boardroom,' which provides pre-built files and coaching to help users implement the technology.

Suggested questions

3 ready-made prompts