How to Get the Open Source Higgsfield AI (Free)
Okay, so today I'm going to show you
guys how to get the open-source
Higgsfield. There are several ways you
can do it, but this is kind of how it
looks when you get it going. This is the
whole system right here. It's pretty
cool. What you can do Okay, so this is
where it all comes from. It's comes from
a Neil Matcha. The jet It's called The
project is called open- generative-ai,
and you have all of the files right
here. So, if you want, you can download
it yourself, and you can edit it however
you want. There are some parts that
don't work perfectly, and that's because
it's an open-source project. That's kind
of how this works. It's like generated
by some person, and they made it for
free, and they're giving this out for
you to use and try. Okay, so there's two
ways that you can use this open-source
Higgsfield. So, we have the pay-per-use
version and the free version. I'll walk
through both of them, and I'll give you
the benefits and the drawbacks. Let's do
this. Okay, so first I'm going to show
you the paid version. So, basically the
paid version is super easy to get. All
you have to do is you go to the hosted
version, which is this website right
here on my screen. You click on that,
and it brings you straight to this thing
that I was showing you earlier. Then
from there, this allows you to use every
single model in existence I can think
of. I I'm sure there's more models out
there, but there's just a ton on here,
and there's lots of options, and it runs
really well because it's running on
their server. But, basically what you do
for this one is it is the paid version,
but it is a lot cheaper than using
something like Higgsfield. Like this is
the open-source version, so this is kind
of like you have like the tiers of
Higgsfield AI is very expensive, then
you have middle, which is kind of middle
of the road, which is this one, then you
have the fully open-source where you
download on your own computer, and that
is free, basically. So, I'll show you
how the middle tier works, and then
we'll get into the free one. Okay,
middle tier, what you're going to want
to do is you will have to generate an
API key. This is a Mu API. I don't know
how to pronounce this properly.
Basically, what you're going to do is
generate an API key in here, and it's
like a mega API key. And so, it covers
every single model you can think in
existence. A lot of them are cheaper
than doing it through the actual
company. So, that's nice. But, you have
access to tons and tons of models. So,
if you get really good with a certain
model that's super cheap, you can
generate a ton of stuff for a lot
cheaper than it would be if you were to
generate it like with Midjourney or or
other companies that host it on their
servers. So, it's a lot cheaper in that
way. So, it's not completely free, but
it is a lot cheaper, which is pretty
nice. But, basically you have all these
options and how it works is it connects
with this API key. You just generate the
API key. I I loaded $10 on there and
I've been generating a ton of stuff. And
that's when you have this section right
here. So, you're able to basically
generate anything through Nana Banana,
Flex Dev, and there's tons of options,
too. You can generate four at a time.
You can change the aspect ratio. You can
add reference images. There's just a ton
of things that you can do. That's for
the image stuff. Same thing with the
video. You have all the options. You
have like a bajillion options. And you
can even download do the open-source
models. They do host them on their own
server, but the thing I'll show you
later is you can host these models also
on your own computer and these videos
for free. I'll get to that. I'll get to
that. Um there also are several models
that you cannot host on your own
computer, and that's like Seedance and
Kling and VEO. These are all hosted on
big fat servers cuz they take a lot to
generate video. But, you're able to do
all those at a more affordable price
here in the video section. So, you can
change the aspect ratio, change the
time. You can change the quality, add
other images. Like, there's a lot of
stuff you can do. You also have this lip
sync section where you can make it so it
matches up with the voice of a face and
you can throw an audio. Pretty simple.
Then, you have Cinema Studio. This is
where you can generate frame-by-frame
and kind of like build the stuff
together. And then, you have Marketing
Studio, which is where you can build out
ads for it. This is also much more
affordable cuz you can choose the model
and you can choose the variant how you
want it to look, the avatar. Like,
there's a ton of cool options in here.
Then, also in here you do have workflows
where you can do certain things which
are like more specialized. So, you do
social media, e-commerce, fashion,
featured, featured social media, like
all these things are are things that can
help you like build better things and
you can build your own workflows as well
and community stuff which is generated
from the community, which is cool. I
like that it's all kind of centralized
into one system. It's like basically the
community building their own version of
Higgsfield that's kind of all combined
here. So, you don't have to pay for
Higgsfield prices. So, you just pay like
the base price and then everybody
contributes to everybody. It's It's a
pretty cool little system. It's not
perfect though. It's not perfect. I'm
not going to act like this is completely
free cuz it's not. There is a free way
to do it. It's not this one exactly.
But, this is the open-source version.
So, like I guess in the sense of
open-source, like anybody can contribute
to this platform right here and help
build it to make it better, which is
pretty cool. You also do have agents
right here where you can literally just
like like look at this. This is crazy.
Like a veggie talk creator. Like there's
a ton of different agents they can use
to make content, which is pretty cool.
All of these are you will have to pay
for the the tokens, but then you also
have apps where you can literally just
build out your own template and create
your own, I guess, wrapper in the sense
of like you can create your own AI video
studio and sell that to other people,
which is pretty cool. And this one you
can just like straight-up go to the
GitHub and do it from there and build it
out. And this is completely free. You
would have to pay for the tokens, but if
you're charging somebody like twice as
much for the tokens, then it can make
sense and it can be an actual business,
which is pretty cool. All these GitHubs
are right there and you can grab the
templates and just download them and
create your own AI business just like
that. It's pretty gnarly how quick and
easy it is. That is the main stuff on
the hosted version, all right? Now, I
know you guys didn't come here for the
hosted version. I just want to show that
this is kind of what the creator is
making. It's like a cheaper and
open-source alternative to doing
Higgsfield, but there is like the other
part where you can just host it on your
own computer. So, let me show you kind
of what that looks like. So, you're
going to go into here, right? Go back to
the main GitHub where I was. And this is
where you can get the fully unrestricted
open-source version of the open-source
Higgsfield. So, so you won't want to go
to this hosted version. You'll want to
download one of these. So, there's
several ways you can do this. For one,
you can just hit this DMG file, which
literally just downloads it like an app.
Like literally just click the button. I
already downloaded it. I don't want to
download it again. It is a decent amount
of memory, but you just hit this button
right here and you just base it on which
type of computer you have. So, I have a
M3 chip, so you just hit this and it'll
download it. Or you have Mac OS,
Windows, um Linux. So, all these things,
these are different versions you'll
click on. Or you can also install it
from the terminal. You'll just run these
prompts in the terminal and it downloads
it. Super quick and easy. Or you can do
the way which I did and I went on the
GitHub desktop app and I just cloned the
repository. Um literally, it's super
simple. You just go on the app, paste
the link and say clone it. And then now
I am editing it in Visual Studio Code
cuz there are some bugs that I wanted to
fix myself that I've been fixing for
certain models. And so, I've been
running it myself. I edited the parts I
wanted to fix in Visual Studio Code so I
can make it better. So, I made my own
little branch. So, you can do that as
well. Um it's a little more technical
and I have Claude Code helping it. So, I
don't know actually crap. I'm not very
technical at all, but Claude Code knows
it and I just tell it like, "Hey, fix
this part. Fix the loading part. Hey,
the open source model's not working
properly. Fix this." So, there's some
things that I did have to fix and I had
to cater it to my computer in order for
it to work properly. So, once I did
that, I'm running it now and it's in
this app Electron right here. And here
is the fully open source version of
that. Okay? So, now there is an option
to run the API key um or you can do the
local models. As you can see, it is a
little bit more laggy, but here are the
models that I'm running right now. So, I
have Z image turbo, which is a image
generator. I've done a downloaded
Dreamshaper and there's a couple other
ones I haven't downloaded yet. When I
When I do with my variations that I did
in Visual Studio Code, it depends on
what computer you have if you want
certain models to work. So, the only
ones that worked on my computer were
these two. So, they're not perfect, but
um they're pretty low-weight models
because my computer's not the highest
end. So, I downloaded these models and I
was able to generate videos, which is
pretty cool. I'll kind of show you what
I generated, but these are the models
I'm running and literally all you do is
you just hit download and this is how
you download them. Some of them don't
work fully and these have to be debugged
and this is just comes with open-source
stuff. It just does. I hate to say it.
So, yeah, image stuff, if you want to go
to your local image generator, hit this
button right here and it'll pick Z image
turbo. You can also change which one you
want and so I have DreamShaper image
turbo. These are all the ones that you
can download on your computer, the
open-source ones, or you can go back to
the API ones and this is charges it to
that same API. But we don't want that,
obviously. We want the free version. So,
we're going to hit this button and we'll
generate images. Now, it does take
anywhere from like maybe like 1 to 3
minutes to generate an image. So, I
pre-generated a bunch of them to show
you kind of what they look like. So,
this is one right here. This is just
like a backyard that I had it generate
super quickly. This is a car in space.
This is some dude in a fake Iron Man
suit. Um this is Elon Musk in space,
which is um it's not bad. And then this
is a ripped frog. This one actually
didn't work. Um and then this one is a
cat on Mount Everest. So, these are ones
that you can generate with your
open-source Higgs field version. Super
simple. Hit this button, type in your
prompt, then you hit generate and it'll
take a couple seconds to load. Now, I'm
not going to generate one right now
because it is pretty intensive and since
I'm screen recording and doing all this
at once, my computer has 18 gigs of RAM
and it'll probably start getting laggy.
I just want to be cleaner for the demo.
Then, we have the video section. So, the
video section is a little more heavy
than the image section. So, since like I
told you earlier, the only ones that my
computer can actually run are has all
these options, as you can see. Oh, it's
lagging. Um is these two, the LTX video
and the Cog video X2B. And it can make
like maybe 2 to 5-second videos, but
here's some of the results that that I
got from what we were dealing with. So,
like this is the type of stuff that you
could generate, depending on how good
your computer is. Like this one's This
one's pretty cool, not going to lie.
Dang, look at that. Car scene. We have
an elephant jumping into some water.
This one's low resolution, so it's a
little bit easier to generate. This is
another car. Looks like it's going to
jump off of a cliff.
Oh, there it goes. And this is a Looks
like a knife on a rock. This is a ninja
right here.
Okay. This is a car through space. So,
like all these are things that you can
generate. The problem with the
open-source models is like basically
what it means is you are running this AI
model on your computer. And that's how
it's free. It's using your hardware and
not somebody else's hardware. If you're
generating with tokens, you're borrowing
their hardware to generate your video.
And that's why you have to pay a little
bit for it. So, um I was doing a bunch
of research and trying to figure out
which models you can run on which type
of hardware. And I built this graphic
that helped me.
Um so, these are all the main
open-source models, and they're pretty
good. They can make pretty good results.
They're definitely not like as perfect
as like the VO models or the Sora
models. But, these ones are pretty good
that you can run. And this is kind of
how it breaks out. So, you got 1 2.1, 1
2.2, High and Video, High and Video 1.5,
um LTX Video, Cog Video, and Cog Video
5B. So, I have an M3 Pro right here. And
so, the two that I'm mainly running are
the LTX Video and the Cog Video X2B. And
so, I could run them. It's not perfect.
Um I haven't tried the 1 ones. Um
they're a little bit bigger, so I was a
little more reluctant. But, depending on
like if you have a basic laptop, I'm I
hate to break it to you, it's just not
going to be Your Your Your computer's
not going to have enough processing
power to be able to run an open-source
model. An Intel Mac, neither, cuz those
are kind of old. M1, you possibly could
do it. My version, I can run it. It's
not perfect, but you can do it. M4 Mac
Pro, if you get a new one, yeah, you
should be able to run these ones pretty
well. And if you have like the top-tier
Macs, you should be able to do most of
them pretty easily. Then we go to the
Windows side. This is kind of like the
GPU you'll need and also the memory. And
one thing that's nice about the Windows
side of things, they're a lot more
optimized for models like this, which is
pretty nice. So, looks like the lowest
end you can run a couple of them, but
you there's a few that you can't. And
yeah, as you see we go up in obviously
price and just performance, this is how
you'll be able to run these things. So,
you'll need like the minimum of this
amount of RAM and this type of processor
if you want to run every single one of
these. The cool part though is if you
have this, you can run all these
completely free and locally, which is
awesome. What you'd basically do is you
would just go in here
and you pick your model, like I pick
mine, and then you just prompt it. And
then from there it's literally the same
thing and you're able to generate all
these same videos with that same model.
Now, I'm I am limited to the lower-end
models because I don't have as powerful
of a computer. And that's just how it
goes with working with ups and
open-source models. You also do have the
lip-sync function, same thing. You have
the cinema studio, this one where you
can like build out frames and use these
to generate cinema cinema quality videos
that you can add into the video function
and the image function.
Unfortunately, the workflows and the
agent functions do not work on the host
the self-hosted version yet, but this is
one thing that you could go into your
Visual Studio Code and edit yourself.
And then you have the MCP 2. This is the
full open-source Higgs Field. It's
definitely not a perfect solution if you
have a lower-end computer, but if you
have a decently powerful computer, you
can run all these things without a
problem. You can run the higher-end
models like I showed you on here. So,
yeah, if you have a higher-end computer,
like one with this type of processing
power, you're you're able to run these
higher-end video models, but if you have
somewhere on the lower end, it's just
it's just not viable. Your computer
simply just cannot handle a model like
this. and I and I hate to say it but
it's just the truth of the matter. And
so if you were in this area down here,
I would probably there's lots of like
technically paid softwares that have
free video credits. So you can do maybe
one or two videos free every single day.
I know Google does it, Gemini chat does
it, Grok does it. There's lots of models
that you can create free video. You just
can't do unlimited and there's certain
limitations. But that would be more than
trying to use an open source model
because you just you simply can't. It's
just your computer is just not powerful
enough. And so that's just how it works
with open source models and this is the
reality of the situation. But if you
want to try out this open source model
for yourself, you have those the ability
to do it. By all means go and do it and
it'll be completely free. But yeah,
anyways, this is the open source Higgs
Field. If you have questions leave them
in the comments, but um yeah.
More transcripts
Explore other videos transcribed with YouTLDR.

O Segredo Que Faz Time QUERER Bater Meta | Conversa de CEO 113
EAG Educação · Portuguese (Portugal, Brazil)

Karier di Bidang Industri Kuliner (Dasar-Dasar Kuliner) SMK Kuliner
Masakmas · English

NR449 Mod4 Nov24
Lisa South · English

Asfixia Perinatal
MayGhis Guizada · Spanish

JUARA 1 PIDATO SUNDA BANTEN JENJANG SD, FTBI SEPROVINSI BANTEN DAN DKI JAKARTA TAHUN 2024
Kantor Bahasa Provinsi Banten · Indonesian

School Tiffin Boxes🍱 For Kids - Leakproof 💦 Testing Done | Best Lunch Boxes From Amazon ✨
Home 'n' Much More · English

How to Get Higgsfield AI Pro Plan 100% FREE! | Unlock SORA 2 & VEO 3.1 Unlimited
NiDat AI · English

PENCIPTAAN JAGAD RAYA DALAM PERSPEKTIF AL-QUR'AN DAN SAINS
Lajnah Kemenag · English

A1 English Listening Practice - Weather
Listening Time · English

RICHARD RASMUSSEN & BIÓLOGO HENRIQUE no FALA GLAUBER PODCAST
Fala Glauber Podcast · Portuguese (Portugal, Brazil)

RICHARD RASMUSSEN e BIÓLOGO HENRIQUE: REVELAM O LADO OBSCURO DA NATUREZA!
O Planeta Felino · Portuguese (Portugal, Brazil)

Warum ich Crimson Desert abgebrochen habe..
gamekritik_ · English
Get the TLDR of any YouTube video
Transcribe, summarize, and repurpose videos in 125+ languages — free, no signup required.