Practice remote viewing honestly.

The honest-measurement problem

You can't practice remote viewing alone honestly — the moment you know the target, you're imagining, not viewing. Every solo practice tool has the same hole: nothing stops you from peeking, and nothing stops you from keeping the hits and quietly dropping the misses.

Aperture closes that hole in code. It plays the blind tasker, walks you through the classic staged protocol without ever showing you the target, and then scores trait overlap instead of the label. Write "cold, white, roaring, falling" for a waterfall and you score a hit — without the word "waterfall." The app rewards what you actually perceived, not a lucky guess at the name.

How a session works

Five staged steps, blind the whole way. A running Signal vs AOL ledger lets you flag any conclusion your mind manufactures and set it aside. The coordinate you see is generated independently of the target, so it can never encode the answer.

1. 00
   Cool-down

   A short breathing pacer to quiet the analytic mind before you begin.

2. 01
   Ideogram

   A fast reflexive mark, read only for gestalt — water, land, structure, motion, life, energy, air.

3. 02
   Sensory chips

   Single words across eight categories. No stories, no conclusions.

4. 03
   Sketch & assemble

   Fuller impressions — drawn and gathered into one picture.

5. 04
   Reveal

   The target appears and the run is scored. Reveal is one-way — cancel before it and nothing saves.

Why blind is enforced in code

The app is the blind tasker, so you never hold the target. Reveal is one-way: cancel before it and nothing saves — you can't keep hits and quietly drop misses. And any analytic guess ("it's the Eiffel Tower") goes in an AOL column that's excluded from scoring entirely, so flagging a guess costs nothing and hiding one gains nothing.

How the score is made

A deterministic on-device score — free and offline — rewards trait overlap between your terms and the target's real traits, with recall weighted over precision: catch the real features first. You see a signal percentage with a Strong / Partial / Low band and a map of hits, misses, and noise.

Over time, Pro opens a longer read of your own record: a signal trend, a per-gestalt profile of what you reliably catch versus miss, a calibration read on your gut rating, and a z-test — are you beating what random term-picking would score? It stays silent until you have enough sessions, and it's heavily caveated. A negative result is the instrument working, not a verdict on you.

AI — a judge, not a genie

An optional on-device model reads your sealed session against the target you were blind to and hands back what matched, what didn't, and one thing to sharpen. Its displayed score is capped so a generous small model can't inflate the honest number, and it never overwrites it. It doesn't tell you what you'll perceive next — it only grades the run you already sealed. Pro, advisory, on supported devices.

What's inside

• ◆
  The blind staged protocol

  Cool-down, ideogram, sensory, sketch & assemble, reveal — the app holds the target the whole way.

• ◆
  Signal vs AOL ledger

  Flag manufactured conclusions and set them aside; flagged guesses never touch the score.

• ◆
  Deterministic signal score Free · offline

  Trait-overlap scoring that rewards perception over naming, with a hit / miss / noise map.

• ◆
  Trends & calibration Pro

  Signal trend, a per-gestalt catch-vs-miss profile, a gut-rating calibration read, and a silent-until-ready z-test against chance.

• ◆
  The judge's read Pro · advisory

  An on-device model grades the run you already sealed — matched, missed, one thing to sharpen — capped by your own honest score.

• ◆
  Relay mode, profiles & export Pro

  Two-person relay on the same blind target, multiple viewer profiles, JSON export, and iPad split-view.

Privacy