Digital SAT, end to end

The whole SAT programme, in one system that actually knows your students.

Classes, an original question bank, adaptive testing and analytics that lead somewhere. Built for departments that are tired of a spreadsheet, three tools and a number nobody can explain.

Start free I have a class code

Every number: traces to a response
Every boundary: enforced in the database
Every flag: shows its evidence
Every rule: editable by you

Four things it does properly

Not a feature list. These are the four places where an SAT programme usually loses information, and what this platform does about each of them.

A question bank that knows itself

Every item is classified by domain, skill and difficulty, and carries the statistics it has earned: how many students answered it, how many got it right, how long it took, and which distractor pulled them away. An item that everybody misses is either genuinely hard or quietly broken — the bank tells you which.

Domain, skill and sub-skill taxonomy
Per-item difficulty and discrimination
Bulk import with a validation preview
LaTeX maths that renders identically everywhere

Testing that behaves like the real thing

Timed modules, a question navigator, mark-for-review, answer elimination, a calculator and a reference sheet — and adaptive routing where module two follows from module one by rules you can read and edit, not by a black box.

Per-module timers with autosave
Explicit, editable adaptive routing rules
Server-side grading — the client never scores
Full keyboard navigation

Analytics that end in an action

Start at a domain, open a skill, open a student, and you land on the individual responses that produced the number. Mastery is a recency-weighted estimate, not a lifetime average, so a student who has turned a corner stops being judged by September.

Drill-down from domain to single response
Configurable mastery model
Intervention flags with their evidence attached
Parent-facing reports in plain language

Built for a department, not a single tutor

Organisations, branches, departments, classes and counselors, with row-level security enforcing every boundary in the database itself. A teacher sees their students. A counselor sees their branch. Nobody sees anything by accident.

Multi-branch hierarchy
Role-based access enforced in Postgres
Append-only audit log
Plan limits held in data, not code

For teachers

The morning question, answered before you ask it

Who is behind, on what, and what should I do about it today. Your dashboard opens on exactly that: the students who have crossed a threshold, the skills your class is weakest on, and the work that is due. Nothing is a vanity metric.

Set a diagnostic to a whole class in one action
Auto-grading the moment a student submits
Build a test by hand or from a blueprint you describe
Export a roster, a report, or the raw responses

For students

A plan, not a pile of practice

After the diagnostic, every student gets a week-by-week roadmap built from their own weakest skills and their own test date — and a review screen that explains the question they missed rather than just marking it red.

A score estimate that says where it came from
Recommendations ranked by what will move the needle
Mistakes reviewed with a worked explanation
Progress measured against a target they set

How it works

01
Create a class
Share a six-character code. Students enrol themselves; you approve nothing.
02
Run the diagnostic
One assignment gives every student a baseline, a skill profile and a study plan.
03
Act on what it shows
The weakest skills are ranked for you. Assign targeted practice in two clicks.
04
Watch it move
Every subsequent result updates mastery, recommendations and the flags automatically.

A student cannot read the answer key. Not through the app, not around it.

Access control lives in the database, not in the interface. Question rows, answer choices and explanations are invisible to a student account; test delivery runs through a function that omits the key; and scores are written by the grading engine alone — a student writing to their own attempt cannot touch a single score column. These are not conventions. They are constraints, and the test suite proves them on every build.

Questions we get asked

Where does the question content come from?: The platform ships with an original bank written for it, and your department can author, import and review its own. No published or copyrighted test material is included or reproduced.
Is this affiliated with the College Board?: No. MEGA Prep is an independent product. It is not affiliated with, endorsed by, or connected to the College Board in any way.
How are scaled scores calculated?: From a conversion table your organisation supplies. Until one is loaded, the platform reports a documented approximation and labels it as such — and it refuses to scale an assessment too short to carry a 200–800 score, reporting a percentage instead.
Can students see the answer key?: No. Question rows, answer choices and explanations are unreadable to a student account at the database level. Test delivery goes through a function that omits the key entirely, and explanations only appear after grading, when the teacher has released results.
What happens to our data if we leave?: It is a standard PostgreSQL database you can export in full at any time. There is no proprietary storage format.

Start with one class. See what a week of real data looks like.

The free plan covers one class and thirty students, with the full question bank, testing engine and analytics. No card.

Create your workspace Compare plans