Tyler's TeardownOne person, one API, a million a year
A bootstrapped image API doing roughly $1M a year, and the one door it left wide open.
Every so often you find a business that makes you slightly uncomfortable, because it is one person doing the kind of number most funded startups never get anywhere near. Bannerbear is one of those.
If you have never run into it, Bannerbear is basically a way to make images without actually making images. You design one template, then it takes some text and a photo and hands you back a finished, on-brand graphic. People use it to pump out social posts, dynamic link previews, and thousands of personalised images, all without a designer ever opening a file. It does video too.bannerbear.com
The money is the wild part. It crossed a million dollars a year this past September, which lands somewhere around eighty grand a month. The year before that it did just under a million from fewer than six hundred customers. The reason I trust those figures, and this part matters to me, is that the founder, Jon Yongfook, has posted his numbers in public the whole way up. No guessing on my end.arrfounder.com+1
How he got there is worth sitting with for a second, because it is almost annoyingly repeatable. It really came down to two moves. He wrote developer content for years, all those quiet guides and integration pages, so the moment someone googled their problem he was already there holding the answer. And he built the whole thing out loud on Twitter and Indie Hackers, roughly half his time coding and half telling people about it, from very early on. The product even started life as a tiny Shopify app before he rebuilt it as an API so it could plug into everything.bannerbear.com+1
So, could you build this yourself? Honestly, yeah. The heart of it, a template editor and something that drops text and photos into a layout and gives you back a PNG, is a few weeks for a decent solo dev, less if you lean on what is around now. The building was never the hard part here. It almost never is.
And here is the thing. Even knowing all of that, I still would not run at it, and you probably want to know why. It is because the obvious version of this is already a knife fight. Placid, Abyssale, Templated, Orshot, there is a whole pile of them, and one literally puts “2x cheaper than Bannerbear” right in its marketing. On top of that, Bannerbear has quietly stacked up two walls you cannot just stroll through: all that SEO you would somehow have to out-write, and the plain fact that once a team has wired their templates into one image API, nobody wants the headache of ripping it back out.orshot.com
But there is a crack in it, and once you see it you cannot unsee it. Bannerbear sells an API, and the people who need automated images the most, a lot of them cannot write a single line of code.
Bannerbear sells an API. But the people who need automated images the most cannot write a single line of code.
Look at who Bannerbear is actually built for and the answer is developers, top to bottom. The pricing is counted in API credits. The docs assume you can hit an endpoint. Even the free tier is capped at thirty images, which is a developer's trial size, not a marketer's. Everything about the product quietly says the same thing: you should already know how to code.bannerbear.com
Now look at who actually needs a steady stream of on brand images. The Etsy seller making a fresh sale graphic every week. The course creator pumping out lesson thumbnails. The agency account manager who needs forty versions of the same ad by Friday. Not one of them is going to call an API. So they do it by hand in Canva instead, the same graphic over and over, because the automatic option was never built with them in mind. There is a whole crowd sitting right next to this product, doing manually the exact thing the product automates, and they cannot reach it.
That is the wedge, and it is not another image API. It is the no code version of one, aimed at a single one of those crowds. Pick Shopify sellers. Connect their store, pull their products in automatically, and generate the sale and product images for them, with not a single template field or endpoint anywhere in sight. Then sell it where those people already gather, the Shopify app store, the e-commerce subreddits, the Facebook groups, not the developer Google where Bannerbear already owns every result.
And the reason Bannerbear cannot just turn around and chase you there is that its whole identity is the developer's API. The pricing, the positioning, everything search has ever learned about it, all of it points at coders. The day it tries to also be the friendly no code thing for marketers, it blurs that story and starts eating into the technical customers who pay it the most. A lean solo team is not going to splinter itself across a dozen little verticals either. That gap between what they are and what this crowd needs is the room they leave you.
So clone the engine if you like, it is a weekend's work. Just do not clone the positioning. The money here was never in another image API. It is in the people the image API quietly forgot.