Yea just skip the ads and go straight to a subscription. It looks like garbage click farm sites. Ads are horrible and tech people disproportionately hate them. So if your goal is to make me think "is there an API?" then you can't look like that.
Look at Stripe/Plaid/etc to get an idea of what an API first sort of site would look like.
The big headline of plaid is "The easiest way for people to connect their financial accounts to an app" and the subtext is "get API key. click here". Don't bury your value proposition behind ads, especially when the ads google will place on your site are (a) competition or (b) spammy "CLICK HERE convert pdf now" type ads.
Visit your site via a public VPN on a clean browser without any google affiliation logged in or cookie'd. See what kind of ads actually show up. Do you want those on your site?
How much is the per-view income from a site that does that? This explains why a lot of crappy sites look like that - handing over control to google to maximize ad revenue income causes google algos to optimize to some maxima of low-denominator quality. Who would intentionally place full-screen ads, after all!?
The main feedback is the visual design of the site is sufficiently non-professional looking (mid 1990's looking?) as to discourage any deeper evaluation of the product. I didn't stick around long enough to assess the product because your marketing materials communicated to this potential customer that your site was not worth looking at in more depth.
Marketing is hard and as specialized a skill as programming. The site at present radiates it's either an ancient or hobby site rather than a service provider customers can depend on. The lack of a visible business model makes the site seem even sketchier. Someone is paying for those compute cycles, so is this site in operation with the goal of monetizing the uploaded data? (As the saying goes, if you can't tell who is paying for the product, you are the product).
There were enough questions raised by that initial quick look at the site to lead me to immediately move on to other competing OCR providers.
As is the case with most tech startups, building the tech is just one of many hard parts of actually doing the startup.