We built the tool we wished our students had.

Why Forge exists

Spend enough time tutoring the SAT and you notice something the score report never shows: most lost points aren't lost because a student doesn't know the material. They're lost to how the student works under pressure. They misread what the question is actually asking. They lock onto a vivid detail and miss the author's real point. They rush the last step, or talk themselves out of a right answer on the second look.

Those are behavioral problems, and they're invisible to a tool that only knows what you got right and wrong. A question bank can't see them. A chatbot that explains one problem at a time can't see them. So we built something that watches the work itself, and turns it into a picture of how you, specifically, take the test.

How the diagnostic works

As you move through the diagnostic, Forge pays attention to the signals around each answer: where you slow down relative to your own pace, when you change an answer and whether you catch your own mistakes, which questions you flag and come back to. It reads that across five recurring dimensions of test-taking thinking, Frame, Observe, Reason, Gather, and Execute, the kinds of moves every SAT question quietly demands.

The score itself is built with Rasch / IRT scaling, the same family of psychometric models the real exam uses, so your number reflects your ability rather than which form you happened to draw. It's an estimate, not an official score, and it's a model we keep refining as more students use it. We'd rather tell you that plainly than dress it up as something it isn't.

Who builds it

Forge is made by Vindica Inc., working with classroom instructors who have prepared students for this test for years. Our founder, Kevin Choi, is an SAT instructor and the author of the Irreverent Guide to the Digital SAT series, on Reading & Writing and on vocabulary. The judgment baked into Forge, what counts as a trap, what a misread looks like, which habits actually move scores, comes from real teaching, not a generic content scrape.

Where the questions come from

Behind the diagnostic is a bank of 4,000+ original and licensed questions, written and curated by instructors and checked for the things that quietly ruin a practice question: a second defensible answer, a missing figure, a key that's subtly wrong. It is not an AI firehose of plausible-looking items. The point of Forge is the read on you; the questions are the instrument, and they have to be clean for the instrument to mean anything.

Your data

Privacy is built into how the product works, not bolted on. Your profile is yours, and it's deletable in one click, no support ticket. When you fold in results from an outside practice test, Forge extracts only the cognitive signal and keeps nothing about the test itself, no questions, no passages, no copyrighted content. The thing worth keeping is the picture of how you think; everything else, we let go.

What we won't promise

We won't promise you a number. Score gains vary by student and by what's actually causing the misses, and anyone who guarantees you points is selling you the fine print. What we will do is tell you, honestly and specifically, how you're testing right now and what's worth changing next, and tell you plainly when we're not yet sure about something. Forge is in open beta; it gets sharper every week, and we'd rather earn your trust than borrow it.