Ask most AI assistants whether your essay is any good and the first thing you'll hear is what's great about it. That's no coincidence. They're trained partly on what people rate highly, people rate agreement and praise highly, and so over time the models learn to tell you what you want to hear. HAITE is built the other way round.
The biggest problem comes first. Not in paragraph four, after the compliments.
It won't reassure you for the sake of it. If you're fishing for "it's fine" you'll get the truth instead, and if the work really is fine it'll say that too, once, without the paragraph of praise first, and then get on with whatever is still holding it back from being good.
Ask for a score out of 10 and you get a number that means something, with a line saying why. A 4 is a 4.
When it doesn't know, it says so. It can search the web and read the links you send it, but if it can't check something it tells you rather than making up an answer.
It's honest about itself too. HAITE runs open models from several AI labs under its own names. It didn't train them, and it'll tell you that if you ask.
Essays and assignments. Paste the work with the marking criteria and you get the weak arguments, the vague sentences and whatever a marker will circle first. There's an AI checker as well, which estimates whether your writing reads as machine-written. It's an estimate, nothing more.
Code. What's broken, what's going to break, and what a reviewer would send back.
Plans and side hustles. The assumptions nobody has tested and the numbers that don't add up, ideally before you've quit anything.
TikToks and YouTube videos. Send the link. HAITE reads the caption, looks at the cover, reads the spoken words when they're available and tells you why it will or won't land.
The honesty doesn't change. The harshness does, because there's a dial from 1 to 10 that sets how much comfort HAITE holds back, starting flat and a bit tired at 1 and ending at 10, where it swears in every sentence and gets personal about what you've made (never about you). If that's what you're after, there's a page on brutal AI.
Because people reward it. Ratings from users feed into how chat assistants are trained, praise tends to get rated well, and the models pick up the habit.
It can estimate it. The AI checker looks at things like sentence length and stock phrases and gives a likelihood. It can't prove anything either way.
Yes. The free plan gives you 42,000 tokens a day. The paid plans add longer replies, bigger models, documents, pictures and video.
No. It's an AI, and it can be confidently wrong about facts, numbers and quotes, so check anything that matters. Treat it as a blunt second opinion, not professional advice.