Prediction-Free

What "Prediction-Free" means

A structural commitment, not a compliance checkbox.

Most apps build a model of you. They watch what you click, infer what you're likely to do next, score how engaged you are, and use that model to steer what you see. The structural fact of being predicted is the problem — separate from whether your data leaks, separate from whether they show you ads.

The Prediction-Free badge says this app doesn't do that. It's a claim about how the product is built, auditable in the database schema and the source code. When this badge is shown, the following seven things are true:

  1. No behavioral prediction. No propensity score, no engagement model, no next-best-action recommender, no churn predictor. Even when the underlying data could support prediction, the product doesn't.
  2. No user modeling beyond the session. Per-user state stops at functional preferences (theme, locale, saved searches, opt-in alerts). No profile of your interests, beliefs, or behavior tendencies is stored for later retrieval or marketing.
  3. No ad targeting. No ads of any kind — first-party, third-party, sponsored content, or affiliate links presented as native content. Revenue model is subscription, B2B, or one-time purchase only.
  4. No third-party trackers. No Google Analytics, Mixpanel, Amplitude, Segment, Heap, Hotjar, Facebook Pixel, TikTok Pixel, or any equivalent. First-party self-hosted page-view counting that stores only anonymous aggregates is allowed, and is disclosed on the privacy page.
  5. No PII handoff to LLM vendors. If the app uses an LLM, your personally identifying information does not leave Modelo Tech's infrastructure. Either the model runs locally, or external calls scrub identifiers before egress.
  6. Transparent retention. The privacy page lists every PII field the app collects, how long it's retained, and how you delete it. No catch-all "we may use your data to improve our services" language.
  7. No silent updates to the claim. If a feature is added that would break any of the above, the badge is removed before the feature ships.

What this badge does NOT claim

  • It is not an end-to-end encryption claim — that's a separate technical guarantee that applies only where actually implemented.
  • It is not a HIPAA, GDPR, or SOC 2 attestation — those are independent regulatory postures with their own audit requirements.
  • It is not "no data collection at all" — the app still collects functional data (account info if you registered, error logs, anonymous page-view aggregates).

How to verify

The seven criteria above are checkable in the codebase. Modelo Tech runs the audit checklist before applying this badge to any app and re-runs it any time the feature set changes materially. If you believe an app showing this badge is violating one of the seven points, email [email protected] with the specific claim and we will investigate, fix, or remove the badge.

Why this exists

Compliance attestations ("we comply with FERPA / HIPAA / GDPR") are paperwork postures, not structural claims. A vendor can comply with all three and still build a sophisticated behavioral model of every user. The Prediction-Free badge fills that gap with a checkable structural claim — the kind of statement a buyer can actually verify against the schema and the source code.

Spec last updated: 2026-05-23.