Master 97% of Canva in 12 Minutes (Beginner to Pro)

Design with Canva (Pro)
19 followers
AI Generated·May contain errors
To answer that question, I went through the whole platform, every main feature and AI tool,
and I built one real brand the whole way through. And in this video, I'll walk you through all of
it, so by the end, you'll know exactly what this tool can do and how to actually use it.
And it starts with the same choice every time, because there are two ways to get a design on
screen and they're not equally good. All of this happens inside Canva and it runs in your browser,
so if you want to follow along, there's a link in the description.
From the homepage, I'll hit the Create button, then look up YouTube Thumbnail,
which opens a blank canvas at 1280x720 with everything you need down the left panel,
including Templates, Elements, Text, Brand, Uploads, Tools, Projects, and Apps.
Templates is the first one, and it's where you'd normally start, because you search whatever you're
making, click one, and it opens straight into the editor with the text, the images, and the colors
all editable. And that last part is where it usually goes wrong, because a template only
stops looking like a template once you've changed all three. There's a second way in though, and
it's the one that honestly sounds better than it is. Back in the left sidebar, Canva AI sits right
at the top, and that's where you can generate a design straight from a prompt. So I'll open it up
and paste the prompt in. The prompt just describes the thumbnail I want, in yellow and in English,
with the headline on the left, the person on the right, and a neon Tokyo street behind them.
And you can find that prompt, along with every other one in this video and the whole workflow,
down in the description completely free, so you can follow along in real time. The thumbnail gets
added straight onto the canvas, and for a first try, it's actually not bad. The layout is bold,
the yellow reads clearly against the street, and the traveler and the ramen look solid. It didn't
follow the whole prompt though, because the headline stops at $10 a day and leaves out in
Tokyo. The text doesn't leave much room around the edges, and the street behind him isn't as
blurred as I asked. So I'll tell it to go again, with the full title, clear space around the text,
and a slightly softer background. This time, the full title is there and the street is blurred,
but the headline still runs right to the edges of the box, so it fixed two out of the three.
So this tool can get you something usable, it just doesn't always follow every instruction,
and it doesn't compete with the top image models. On the free plan, you only get 20 AI users a
month, and every retry counts as one. So it's worth trying, but you don't get unlimited attempts.
It does come back as separate editable layers though, the headline, the person,
and the background, so you can at least edit it as is, instead of getting back a normal image.
The tools tab is worth knowing about as well, because that's where all the manual stuff is.
The shapes, lines, sticky notes, texts, signatures and tables, all your standard editing assets.
So even with AI included, templates are still where I'd start, and I'd use the generation tool
to try out different ideas. Either way though, what comes out still looks like canvas design,
rather than yours. And the fix for that is the one thing I'd set up before anything else,
because it changes every design you make after it. That's the brand kit, which holds your logos,
your colors, and your fonts, so every design pulls from the same set instead of you choosing
them over and over. You'd normally build one by hand, adding each color and uploading each logo
one at a time, but Canva will read a brand guidelines PDF and build the whole thing off
that instead. I don't have a brand to test it with, so I'll get Claude to invent one and turn
it into a guidelines PDF. It comes back with 3 directions, and I'll take Nightbloom, a sparkling
soda brewed with flowers that only open at night. Before it builds anything, it writes out the spec,
the tagline, the logo, 5 colors with their hex codes, 3 fonts, and the brand voice. Then the
document comes back as 6 pages, with the primary logo, the reversed logo, and the mark attached
separately as transparent PNGs. Back in Canva, I'll open brand from the homepage, where there's
a banner that says get your brand kit started in seconds, so upload your PDF, pick the guidelines
file, and that's it. Now it reads the document and builds the kit on its own, named Nightbloom,
with the mark as its icon, all 3 logos in with their guidelines, and every font it recognizes
already assigned. It also takes the named palette out of the document and pulls a separate one from
each logo file, and it splits the brand voice into a do and a don't list. And on the free plan,
a brand kit caps out at 3 colors, which is worth knowing before you build one, since Nightbloom
has 5. So the whole brand is in Canva, and I never entered a single color or font by hand,
but let's see how it holds up when I actually try to design something. The first thing this brand
needs is a launch post, and I'll use a specific tool to design it. That's Canva AI, which is in
the left sidebar, but it's only available on the pro tiers, so if you're a free account, you'll
need to use templates instead. The prompt asks for an Instagram post announcing the launch,
a glowing can in a moonlit garden, and it gives the headline and the line underneath word for
word. The only thing it says about the brand is to use my Nightbloom brand, and that one line is
keeping everything consistent. Canva picks the kit up on its own and already has the brand tab open,
so I'll pick Nightbloom and hit generate. The post opens beside the chat with the full editor sidebar
next to it, and the logo, the palette, and the fonts are all on it. None of that was in the
prompt, and that's the whole reason I start with the brand kit. The headline and the line under it
are live text layers too, and the logo mark is its own element, so it's all still editable,
but one post is rarely enough, because the same design usually has to go out in a few different
sizes. So with it still open, there's a resize button up in the editor menu bar. You get preset
sizes, a browse by category list, or a custom size with the aspect ratio locked or unlocked,
and you can select more than one of them at a time. Then you choose between copy and resize,
which leaves this design alone and saves the new versions into the same folder, and resize this
design which changes the one you're in. I'd use copy and resize every time, because the original
stays exactly as it was. It's a pro feature though, and each resize counts against your
monthly AI usage, so it isn't unlimited even on a paid plan. So one prompt gave me a finished
post with the right brand, and I can instantly resize it for any format I need. That's one image
though, and the moment I asked it for something longer, this stopped going so smoothly. So in
the same chat, I'll ask it for a 6 slide pitch deck, and this is where you have to start checking
its work. The prompt carries the same idea further, a deck pitching night bloom to rooftop bars and
late night cafes, keeping the same look as the post, with a line for each of the 6 slides.
It doesn't build anything yet though, it comes back with a presentation outline first, one card
per slide, and you open each one and read it before anything gets made. All 6 slides come back
holding the same look as the post, so the brand kit is still coming through. Slide 4 is wrong
because all 3 cans spell the brand as night loom, which is a problem. The name is right everywhere
else in the deck, so it only slipped up on the cans themselves. So I'll correct it again and
tell it to fix the name on all 3 cans on slide 4, and keep the text as it is. But it still doesn't
get the name right, since AI image tools still struggle with spelling out text in pictures,
so that's something you'd have to fix yourself before using it anywhere. So even though it's
not perfect, it's still faster than doing it by hand, and the most impressive thing is that
the brand's identity holds up all the way through, which is what keeps everything professional.
So we saw how Canva does when building a design from nothing, but how will it do when I want to
edit a pre-existing image? I need a photo to work on, and while you'd normally just upload your own,
I'll generate one here to walk you through everything. I'll start by opening a new Canva AI
chat and ask for a photo, rather than a design, of a woman on a rooftop bar at night, holding a
nightbloom can. Then I'll select the photo and hit edit in the top toolbar. That opens the panel
every one of these tools runs from, and in there, adjust and filters are the standard photo controls.
Everything else is AI, and the three I'm about to run are all on the pro tiers. First is background
remover, and it's pretty simple, I'll just click it and now everything behind her is gone, but the
table and everything on it remain. Then there's magic eraser, and it gives you four ways to pick
what goes. With brush, you paint over any area, click lets you click an object and it selects it
for you, foreground takes the whole foreground at once, and text erases writing that's part of the
image itself. I'll use click, so I'll select both of the glasses and hit erase, then do the same for
the napkin. One of the glasses and the napkin came out clean, but instead of removing the other glass
entirely, it also left a small patch of color noise where the glass was. So one of the glasses
is a job for magic edit instead, which swaps whatever you select for whatever you describe.
I'll switch select area to click, click the glass, and ask for a plate of loaded nachos.
This one doesn't come out great though, the nachos look fake and don't really blend into the shot,
so you might need a few tries to get something usable. These tools might seem pretty standard,
but they're essential when it comes to editing, and all those tedious tasks you use to handle by hand
take just one button, and on that same panel we can do a lot more. So the next one stops treating
it as a photo altogether. That's magic layers, and one click splits the photo into separate objects,
everything from her and the can down to the lights. So I can drag the can out of her hand
and down onto the table, which you can't usually do to a normal photo. It doesn't come away clean
though, because the hand gets filled in with a smudge. You can try this one on a free plan too,
it just comes out of those 20 uses. Having the layers split out for you is very useful,
just don't expect the actual movement between them to be that great. Then there's background
generator. So this time, instead of erasing what's behind her, I'll just describe what to put there.
The prompt is just a Tokyo street at night with food stalls and steam. The swap itself is solid,
but the lighting doesn't match. She's still lit in purple, while the street behind her is warm
yellow. So this clearly looks edited, but depending on the use case, that's not always an
issue. Upscale is something you can use for most projects, and it only takes a single click,
bringing the photo from about one and a half megapixels up to 15, which is nearly 10 times
the resolution. Her hair, the lights, and the glass edges all come out sharper, though her skin
gets slightly smoothed. And style match redraws the whole photo in a style you pick, like scrapbook
or doodle, with everything staying exactly where it was. So I was able to completely change that
one photo, and all those features work almost instantly. Now, everything so far has happened
inside Canva itself, but you don't actually have to be in there to make any of this. Canva actually
has a connector for Clod, which is very useful when you're managing multiple projects. I'll show
you how to set it up in four simple steps. In Clod, you open Customize, then Connectors. Search
for Canva, hit Connect, and sign into your account to allow access. After that, you just ask for what
you want. So I'll ask for a thumbnail for a video about testing 50 cheap gadgets and tell it to make
it pop. It comes back with three directions first, and once I pick one, it writes the full brief and
shows me that before anything gets sent to Canva. The design comes back into the chat as a design
ready card, and Edit in Canva opens it straight in the editor. It's a lot better than the first
thumbnails we made, with bold standout colors and a playful look that fits the video perfectly. And
it actually followed the brief this time, the red arrow, the cyan circle, and the worth it sticker
are all there and spelled right. And everything in it is still editable, so if anything did come out
wrong, you could just click the text and type over it. Now the finished result isn't as good as
something you design professionally, but it holds up well for something that comes back in just a
few minutes. If you're testing out ideas or want to create content at scale without having it be
top-notch quality, this will get the job done. The same thing works for a logo, so I'll ask for one
for a travel brand called Driftline Travel that has to work as a profile picture. It gives me three
directions. Again, I'll take the wave one and hand it the colors, and it writes the brief. That one
opens in Canva, with the wave, the wordmark, and the background as separate elements, editable like
everything else. This is simple, but it's a lot easier to work with. Once you're done, you can
download your design in multiple formats, and on a free plan, you can choose the size, the quality,
and which pages to include. Transparent backgrounds, SVG, compressed files,
and bulk downloads are all on the pro tiers. And if your design uses any premium elements,
a free account either pays for the license, upgrades, or downloads it with a watermark on it.
So that's basically all of Canva, from a blank screen to a full brand, and most of the work in
there is now something you describe rather than build. And if you want the same level of control
when you're making videos, click here to learn how GPT-6 Astra directs everything for you.
Thanks for watching, and I'll see you in the next one.
