About

Why I built this

Almost everything you read about making money online is nonsense. Dropshipping courses, crypto tips, passive income that turns out to be someone selling a course about passive income. I had written off the whole category.

Then I found the exception. The companies building AI models pay people to train and check that work, and they pay properly for it. A lawyer reviews legal reasoning. A doctor grades medical answers. An engineer reviews code. A native speaker rates translations. It is ordinary skilled work, done remotely, paid by the hour, and the rates are printed on the companies' own job boards. Nobody is selling you anything.

I checked it myself rather than take anyone's word for it. I applied, sat the interviews and assessments, worked on real projects and got paid. The rates are real, and each one is the most a role says it can pay. How many paid hours you get depends on the project. For me it has been paid work on top of a day job, done in the evenings. The projects are not exclusive either, so you can work for more than one platform at a time.

The problem

Not many people know this work exists. The ones who find it hit the second problem: it is scattered. Every lab and every staffing vendor runs its own board, so finding the good roles means keeping a dozen tabs open and checking them again and again.

The third problem is the one that makes people quit, and nobody warns you about it. Your odds on any single role are low. I applied to a lot of them before I was accepted anywhere. Most people try three or four, hear nothing, decide the whole thing is a scam or a waste, and stop. They were not doing badly. They were doing the normal thing at a volume that was never going to work.

What actually works is the opposite: apply early and apply widely, because there is almost no cost to being turned down. You are not spending a reference or a reputation, and nothing carries over to the next application except in your favour. Several platforms now interview you once with an AI and reuse that interview across roles, so the expensive part happens a single time. After that, applying is often one click. Being early matters as well, since good roles do close. But speed is not the whole story: my own replies have ranged from a couple of days to more than a month, and fit and availability decide as much as timing.

What I wanted was to stop finding out about suitable work late, and to stop guessing which of it was worth an evening. That, more than anything, is why I built this.

The sites that try to gather it all up are bad. Listings that closed weeks ago. No pay figures at all, or a “$150/hr” headline that turns out to be a ceiling almost nobody reaches. Rankings that quietly put whoever pays the site at the top.

So I built the one I wanted

One list, every role read straight from each company's own job feed, checked every few hours so that what you see is actually open. Pay is shown as the top of each advertised range, because the top is what the role can pay and what you should know before deciding whether to apply. When a listing publishes no pay, the site says so rather than guess. The methodology page spells out every rule.

I built it for myself first. Once it worked, keeping it to myself seemed silly, so here it is.

The listings tell you what a role pays. They do not tell you what applying is like, so I wrote that down separately, from my own applications: what happens after you apply covers the interviews, the rate you are asked to name, onboarding and when the money actually arrives.

What I leave out, on purpose

An aggregator that lists everything is not doing you a favour. Two kinds of listing are missing here deliberately. Roles advertising less than $10 an hour are not listed, because this is meant to be work worth an evening, and a $4 row sitting next to a $400 one teaches you to trust neither number.

Whole platforms come off too. Before listing one I read what people who actually did the work say about it, and some of them do not survive that. The pattern I look for is a headline rate that belongs to a small specialist tier while the work most people can actually get pays single digits an hour, usually because pay is per task against a time budget rather than by the hour, so working carefully lowers what you earn. Add long unpaid qualification and stretches with nothing in the queue and the advertised number stops meaning anything. Passing that on to you would make every other figure on this site worth less.

When I drop a platform I say which one and why, on its own page, with the sources and the date, because quietly deleting things is the behaviour I built this to get away from. I also write down what is true in its favour, because being left out here is not an accusation of fraud: most of these companies really do pay people. If one starts publishing rates its own contributors recognise, it comes straight back. See which platforms are affected.

How it pays for itself

Some platforms pay a referral bonus. If you sign up through one of those links, pass their process and start earning, I get a small one-off bonus. That is the main business model, and it covers the running costs and the time to keep making this better. I also offer optional coaching for the AI interviews and tests, with no upfront payment and the cost agreed by email.

It buys nobody a better position. Referral payments never affect the order of roles, and the disclosure page names every platform that pays me anything. Most links here earn nothing.

Who I am

I am Alex. I work in operations and strategy in Europe. There is no company behind this, no investors and no team. If something here is wrong, tell me at contact@aipaylist.com and it reaches the person who wrote the code. I read everything.

The site uses page analytics, including Google Analytics cookies, and I do not sell anyone's email. The privacy page is short and says exactly what is stored.

Common questions

Every number on this page is computed from the live index when the page is built, not typed in by hand. See the full pay index →