Built by medical students

Find what you’re missing.

You already do the reviews. Lacuna reads them and finds the concepts you keep getting wrong: the ones you would never have thought to study.

What it finds

The mistakes you don’t know you’re making.

One wrong answer is noise. The same idea missed the same way, three weeks apart, in Anki and again in a case: that is a gap, and it is the kind you cannot see from inside your own study.

Written for your mistake

Not a textbook page. The explanation opens on the part you actually got wrong, and ends with the one sentence to carry into the next question.

Across both your sources

The same concept missed in Anki and in a clinical case counts once, not twice. Cross-source matching is the point of connecting both.

Every result is auditable

You can see how each review was understood: topic, objective, confidence, route. You can correct it, and nothing is presented as mastery or competence.

Explanations

An explanation written for your mistake.

Ask for any concept and you get an explanation shaped around the mistake you actually made, not a textbook page to search through. A disease opens on pathophysiology, a drug on mechanism, and whichever part you got wrong is already open.

Any concept, any system, at the depth the concept deserves. It keeps improving as the people using it correct what they find.

The library

Every concept you study, already covered.

Whatever you study is already here, or will be the moment you study it. The library grows from what learners actually review, so it covers your curriculum rather than someone else’s.

Your library, and everyone’s

See only the concepts your own reviews have touched, or the whole communal library that every learner is improving.

Explanations that stay current

When someone corrects a fact, it is corrected everywhere it appears, so what you read has survived every learner who met it before you.

It gets better as you use it

The library grows from ordinary study rather than curation, so the concepts your cohort actually reviews are the ones that get deepest.

Peer review

Every fact has someone’s name on it.

AI writes the first draft. The people using it correct what is wrong, and approved reviewers put their name to what is right.

AI-written

Thiazides are first-line for hypertension in most patients.

Attested · 3 reviewers

Thiazides are first-line for hypertension in patients over 55, or of Black African or Caribbean origin at any age. ACE inhibitors come first otherwise.

A learner reading this spotted the missing qualifier. One correction, and every explanation built on that fact changed with it.

Auditable by design

Nothing about you is guessed at.

Before Lacuna tells you a concept is a problem, you can see exactly which reviews led it there and how sure it was about each one. Disagree with any of them and say so.

How it works

Study the way you already do.

1Review as usual

An Anki add-on and a browser extension record that a review happened and how it went.

2Each review is mapped

A model assigns the topic and learning objective. You can see and correct every mapping.

3Patterns surface

Versioned, deterministic code decides when a repeat becomes a focus area. Not the model.

4You study smarter

An explanation written for the mistake you actually made, checked by other learners.

Where it runs

No new place to study.

Lacuna works inside the tools you already open. Nothing to migrate, no deck to import.

Plans

Free during the beta.

Plans differ by how much AI analysis you get, how many devices you connect, and how far back your history goes. Analysis on your own device never runs out, and nothing is charged automatically.

Pulse

One learner, one device. Everything stays yours.

25AI analyses / month
1connected device
14 daysof history kept

On-device analysis is unlimited.

Atlas

For programs running Lacuna across a cohort.

5,000AI analyses / month
10connected devices
730 daysof history kept

Adds privacy-preserving cohort analytics.

These allowances are initial beta controls, not final prices.

Privacy

Your cards never leave your device.

Lacuna needs to know that you reviewed something and whether you got it right. It does not need to see it.

What Lacuna receives

That a review happened, the concept it belongs to, and whether you got it right.

What it never receives

Card fronts and backs, note fields, question and answer text, images, screenshots.

Turning on AI analysis sends a concept name and objective, never your material. You switch it on per device, and it works without it.

Start with the reviews you have already done.

Connect Anki or a question bank, keep studying exactly as you do now, and let Lacuna find what keeps going wrong.

Join the free beta