Help center

Interlaza is an adaptive learning platform built around match-to-sample training. The engine presents a learner with a sample and a set of comparisons, tracks every response, and adjusts difficulty one concept at a time.

It is designed for three kinds of people:

• Instructors and families who want a structured, low-friction way to teach early concepts.
• Professionals who need per-learner data, analytics, and remote session supervision.
• Researchers who want raw data export and fine control over experimental parameters.

Interlaza runs in the browser, works offline on a tablet, and syncs data when you come back online.

1. Sign in at interlaza.app.
2. Create a learner profile — just a name is enough to start. Details like school or class can be added later.
3. Pick a pathway from the library (for example Interlaza, Vocabulary, or RFT Foundations) and clone it.
4. Assign it to the learner and hand them the device.
5. Review data afterwards — every response is logged and summarised per concept.

You do not need to configure anything the first time. The defaults are set to be low-pressure and errorless, so the learner succeeds early and the engine gets a clean signal to start adapting.

Three systems work together behind every session:

Errorless fading

At the start, the wrong answers are faded out so the learner almost cannot miss. As correct responses accumulate, the distractors gradually fade in. The goal is to build confidence first and introduce difficulty only when the learner is ready.

Mastery tracking

Each concept has an independent mastery estimate. Instead of a single percentage, the platform uses a probabilistic model (Bayesian Knowledge Tracing) that weighs recent evidence and asks how confident are we that this concept is mastered? Mastery is declared only past a high confidence threshold.

Error pattern detection

While a session runs, the engine watches for patterns: always picking the same position, guessing too fast, repeating the same mistake. When a pattern triggers, the session nudges itself — for example, by balancing positions or slowing the pace — and the instructor sees a quiet alert.

There are four paid plans. What changes between them is not the core learning engine — every learner gets the same adaptive experience — but the capacity and the control surface.

• Family — one learner, template-based pathways, simple review screen.
• Professional — multiple learners, analytics dashboard, remote supervision, AI-generated summaries, webhooks.
• Researcher — up to 25 learners, raw data export, Varela transfer experiments, custom dimensions, full expert mode.
• Institution — everything above plus organisation-level administration. Pricing on request.

Every paid plan starts with a 14-day free trial. If the trial ends without a subscription, the account keeps all data but is switched to a read-only Free tier. You can reactivate any time and pick up exactly where you left off.

Current prices are shown on the pricing page.

Interlaza is offline-first. Sessions, learner profiles, and results are written to your device's local database the moment they happen. When you are online, they sync to a private Supabase project under your account. Nothing is shared with third parties for advertising — we do not sell data.

Exports

From any learner screen you can export results as JSON or CSV. Researcher plans can export the raw trial log (one row per response).

Deletion

From Settings → Privacy you can permanently delete your account. This removes every record in the cloud and triggers a local wipe the next time the device goes online.

Compliance

Interlaza is designed to meet GDPR requirements and — for family accounts with minors — COPPA. The privacy policy has the full details.

For anything not covered here, write to support@interlaza.com. You can also use the contact form, or open the in-app Contact Support panel from the sidebar — that one pre-fills your account email so we can see which learner profile you are asking about.

Response times:

• Billing and access — within one business day.
• Bugs with a clear reproduction — same day if it breaks a core flow.
• Feature requests — we log every one and reply when it lands or when we decide it will not.

If your question is urgent and learner-facing, write URGENT in the subject line and describe what the learner sees.