Not Everything Happens on a Screen
Hybrid activities with and without a tablet: combining materials, recording what happens, and studying transfer without confusing data with conclusions.
Read more For Instructors
You decide what to teach. The app runs the session and writes it down.
Interlaza shows the child what to look for, gives the help you chose and takes it away as they get it right, and notes every answer together with the help that came with it. When you sit down to review, the session is already written up: what the child did alone, where they needed a hand, and what to check next.
Match-to-sample: the child sees one thing and finds it among several · help that fades as they get it right, on a rule you set beforehand · one row per trial, exportable whole
Free for everyone during the testing phase, until 30 November 2026 — no payment required.
What it changes in your week
Less time writing things up, laying out materials and reconstructing what happened — so the review starts with the evidence already in front of you.
- Interlaza programs ready to use, and routes shared by other Instructors that you can edit
- The app keeps an estimate of what the child already knows, trial after trial, so a lucky run of correct answers is not read as learning
- It warns you during the session when the answers fall into a pattern — always the one on the right, leaning on the help — and proposes an adjustment you apply with one tap
- Draft progress reports written from the session data, for you to review and sign
- One history per child: what happens on the tablet, what you run with cards or objects, and what you simply observe
- That estimate helps you decide the pace. Whether it carried over to new material, to another setting, or still held a month later stays a separate check
- Use the Interlaza programs as they come, or build your own from scratch. You choose how much you want to decide.
How It Works
One engine. Every exercise type.
From identity matching to equivalence relations and relational frames — all in one adaptive platform.
Multiple exercise types
Identity matching · Non-identical matching · Symbolic (word→image) · Auditory (sound→image) · Equivalence, relational frames, and more. Each type adapts its difficulty automatically based on the child's real-time performance.
800+ concepts ready to use
Our own library — more than 2,100 images across realistic photographs, pictograms, and scenes, plus geometric shapes and sounds. Created and curated for this platform, organized by category, animacy, size, and frequency. Add your own photos or AI-generated images at any time.
Instructor Mode
Off-tablet sessions, recorded in one tap.
Sometimes the child works with physical cards, real objects, spoken language, or in a setting where a tablet doesn't fit. Interlaza tells the instructor exactly what to present next and how many trials to run, then captures each response from the instructor's phone — independent, prompted (with prompt type), incorrect, or no response. Same trial-by-trial data as a tablet session, with the program doing the planning.
The program plans. The instructor records.
The header shows which card is the sample and which is the target (ringed in amber), plus the trial counter — Interlaza picks the next stimulus and the number of trials based on the child's mastery level. Four zones below — Independent, Prompted (with five prompt types), Incorrect and No Response — let the instructor record each response with one tap, eyes on the child, not on the screen.
For paper, objects, and hands-on teaching
For when the tablet is not the right medium, or the learning needs physical materials in hand. The Instructor keeps the phone, with the recording pad open, and presents the real objects. These sessions land in the same record as tablet ones — one continuous history per student.
Three ways to run a session off the tablet
Matching with cards
Lay out physical pictogram or photo cards from the library and record each match. Interlaza tracks the sample, the target, and the prompt level.
Matching with words only
Define your own samples and comparisons as plain words — no images needed. Ideal for readers, labels, or materials you bring yourself.
Free Operant
No cards to choose from. You give a cue and score what the child says, sorted into four kinds of spoken response: repeating, asking for something, naming, and answering.
Before each session, print the trial plan — sample, comparisons, and targets laid out as a checklist, or pictograms as ready-to-cut cards — so the materials are staged before you sit down with the child.
What the app does
What an Instructor or a researcher needs
- Adaptive MTS Training
- Match-to-sample exercises that get easier or harder as the session runs, following what the child is getting right and wrong.
- Instructor Data Entry
- Run a session with physical materials, off the screen, and score each trial yourself on the recording pad — no tablet in the child's hands required. Live IOA shadow scoring lets a second observer compare in real time.
- Free Operant
- Score spoken or signed verbal behavior off the tablet — echoic, mand, tact, and intraverbal — with no comparison cards. Present a cue, record the response, and that response joins the same mastery and BKT tracking as every other activity.
- Task Chaining
- Teach multi-step routines — hand washing, brushing teeth, getting dressed, making a sandwich, writing a name — forward, backward, or as a total task, with a seven-level most-to-least prompt hierarchy and per-step scoring of independent vs. prompted.
- Direct Measurement
- Not every target is a trial. Count a behavior to get a rate per minute, or time episodes to get total duration and duration per occurrence — the two numbers one twenty-minute episode and twenty one-minute episodes do not share.
- Bayesian Knowledge Tracing
- Each concept carries its own estimate of how likely it is to have been learned, updated after every answer, so mastery is a probability you can look at rather than a percentage you have to interpret.
- Stimulus Equivalence
- Train two relations and test for the four that were never taught. Sidman's equivalence paradigm, with the symmetry and transitivity probes run without feedback.
- Relational Frame Theory
- Coordination, distinction, opposition, and hierarchy frames built in, with contextual cues you can configure per exercise.
- Errorless Learning
- Configurable distractor fading — 3 phases by default — with 5 prompt types: opacity, size, highlight, arrow, and position.
- Smart Interventions
- While the session runs, the app watches for ways of answering that say nothing about what the child knows — always the same side, repeating one answer, answering before looking — and adjusts, showing you what it changed.
- Transfer Matrix
- Following Varela: a concept counts as mastered only once the child is also right on material they never practised, not on the trained material alone.
- Clinical Reports
- AI drafts a progress summary from the session record and points at things worth a look. The Instructor reviews and decides — none of it is a treatment recommendation.
- Open Integrations
- Connect with Zapier, Make.com, Google Sheets, and IEP systems via webhooks. Export data where you need it — no closed ecosystem.
- Community Store
- Browse and clone training modules shared by instructors and researchers. Filter by paradigm and age group, review the author and configuration, then adapt a copy before assigning it.
- Your own images and sounds
- Upload your own photos, drawings, or audio clips and use them in any exercise — bring familiar faces, household objects, or context-specific items into training without losing anything else the app does.
Clinical Analytics
Data that holds up when someone checks it. Every session.
Every trial records phase, prompt, response time and error patterns.
Mastery Curve
An estimate after each trial that counts · with your criterion in view
Session Accuracy
Independent vs prompted · 8 sessions
38%
S1
47%
S2
55%
S3
61%
S4
70%
S5
79%
S6
85%
S7
91%
S8
Independent Prompted
Error Patterns
5 detectors · Every 10 trials
Position bias
18%
Overselectivity
12%
Prompt dependency
74%
Perseveration
8%
Fatigue signal
22%
Prompt dependency detected (74%) — 3-correct fading applied
Generalization Probes
3 levels · No feedback
Level 1 — Intra-exercise
Novel exemplar
94%
Level 2 — Cross-modal
Different modality
78%
Level 3 — Cross-exercise
Untrained exercises
61%
The measurement itself
Before you act on a number, three checks on the number
A percentage on a chart is worth whatever the session behind it is worth. Interlaza runs three checks on that session and shows you what each one found.
Would somebody else have scored it the same way?
Two people watch the same session and mark each answer without seeing what the other marked. Interlaza compares the two records trial by trial and reports how often they agreed — per session and per concept, never as one average over everything — alongside a second figure that sets aside the agreements chance alone would produce. That comparison is interobserver agreement, and it is what turns a score into a measurement a second person could repeat.
Did the teaching run the way it was planned?
The app already records what it did on every trial: which help it gave, how long the pause was, how many options were on screen. Interlaza reads that record back against the plan you set, with nothing extra for you to fill in, and names whatever it could not see instead of folding it into the score. That reading is procedural fidelity — and while it comes back unclear the app holds back its suggestions about the child, because a flat line can be describing the adult.
Which trials count, and which deliberately do not?
Some trials are there to check rather than to teach: a few trials with no help, no feedback and material the child has not practiced. Interlaza keeps those out of the mastery count in both directions — they never award mastery and they never break a streak. A trial the child did not answer is recorded as unanswered, not as an error. And the criterion that decides when a concept counts as learned is one you set in advance, not a number the app picked.
What these three checks do not do: none of them says the child learned. They say the record of the session can be read as the thing it claims to be. Whether what was taught carried over to new material is a separate check, and Interlaza reports it separately.
The reason behind every rule in a session →Plans for professional use
Free for everyone during the testing phase, until 30 November 2026 — no payment required.
Professional
5 students
For professionals managing multiple students.
- Everything in Family
- Up to 5 student profiles
- Ready-made Interlaza programs or create your own
- BKT adaptive learning
- Supervision dashboard
Researcher
Up to 25 students
For researchers and advanced practitioners needing full control.
- Everything in Professional
- Up to 25 student profiles
- Generalization & transfer probes
- Cohort comparison analytics
- Raw & anonymized data export
Institution
Up to 100 students
For clinics, schools, and organizations.
- Everything in Researcher
- Up to 100 student profiles
- Organizational management
- Priority support
- Custom onboarding
Frequently Asked Questions
Is it suitable for school or clinic use?
Yes. The Professional plan is designed for clinical and school settings. The Institution plan adds organizational management features for clinics, schools, and multi-professional teams.
What device do I need?
Interlaza runs in the browser — nothing to install, though you can add it to the home screen and use it like an app, offline included. The child-facing screen scales to whatever size it is given and never shrinks a touch target below 80 px, so a phone works. A 10-inch tablet is the best size: that is where the images render at full scale. A 7–8 inch tablet is comfortable too. Instructors can plan sessions and record off-tablet trials from a phone or computer.
Who owns the clinical data?
You do. All session data is stored locally on your device. Cloud sync is optional. We never sell or share clinical data — it belongs entirely to you and your students.
How does Interlaza handle protected health information (HIPAA)?
Whether HIPAA applies at all depends on you, not on us — it binds covered entities and their business associates, so a private family is outside it and a clinic usually is not. What we can tell you is how the data sits: Interlaza is local-first, so session and student data is stored on your own device rather than on our servers, and cloud backup is optional and off by default. With it off, nothing leaves the device. If you turn it on and your organization handles protected health information, you will need a signed business associate agreement and your own risk analysis — contact us before switching it on. We never sell or share clinical data.
Can I manage more than one student?
The Family plan supports a single learner. Professional includes up to 5 students at a flat monthly price. Researcher covers up to 25 students. Institution offers up to 100 students with custom pricing.
Articles for you
View all articles →Two Questions About the Same Card (II): Ribes' Question, and Where the Two Meet
Ten out of ten says nothing about what kind of contact the child has with the cards. Ribes tells five apart, and Varela's tests are how you find out which.
Read moreTwo Questions About the Same Card (I): Varela's Question, or What Survives When You Change Something
A child picks the blue circle when asked for the same one. Right or wrong? Julio Varela answers by changing one piece and watching whether the answer holds.
Read moreStart with one ten-minute session.
Free until December: we are in testing. No card, no payment.
Start free Or see the plans first