๐Ÿš€ OrcaScope private beta โ€” first 10 design partners get 3 months free Join now!
OrcaScope
Inside the platform

Visibility tracking

The same prompt set, asked across every engine, every day โ€” so you can see whether you're rising or falling.

Cadence

Once daily

Granularity

Per prompt ร— engine cell

Score range

0โ€“100, withheld when data is thin

To start

One brand and a prompt set

The shape of it

This diagrams the concept, not the UI โ€” screenshots go stale on every redesign; structure doesn't.

A matrix of prompts by engines, aggregated into one visibility scoreEngines โ†’Prompts โ†“0โ€“100one point per dayFilled = that engine named you for that prompt

What it does

One score, plus a breakdown per engine

The headline number is for reporting; the per-engine split is where you find what's dragging.

Citation sources

Who the AI cites when answering your category โ€” that's where your next content should go.

Industry percentile

Compared against the anonymized industry distribution, so a 62 actually means something.

Scoring version on every record

Every score carries its version, so old data is never silently recomputed under a new formula.

How to get going

  1. 1

    Register the brand and its aliases

    Aliases matter โ€” if the same brand appears under two spellings, missing one halves your hits.

  2. 2

    Set a category prompt set

    Ask "which X are worth considering", never name yourself โ€” self-naming prompts are excluded from scoring.

  3. 3

    It runs daily from then on

    You'll have your first data point the next morning. A trend takes several days โ€” we won't pretend otherwise.

How it actually works

These live behind a login, so you can't verify them on the spot. That's exactly why the mechanics are spelled out here; each line maps to a specific implementation.

Scores are computed server-side, then sent down

The browser only renders; it never recomputes. Two independent implementations eventually disagree.

Prompts that name you are excluded

If the question names you, the AI will almost always mention you โ€” that measures phrasing, not visibility. We learned it the hard way: 288 of one brand's 296 cells had to be discarded.

Not enough sample means no score

We'd rather show "not enough data" than a number that looks precise but rests on two or three cells.

What it can't do

This section matters as much as the ones above. Leaving it out would decide "this fits me" on your behalf.

  • It won't predict tomorrow's score. Nobody can.
  • It won't promise a change will lift you; it honestly records what happened after you changed something.
  • It can't see your site traffic. We install nothing, and don't intend to.

Other capabilities

Want real numbers before signing up?

The free checkup needs no account and shows you in 30 seconds how AI answers your category today.

Run a free checkup
Visibility tracking ยท OrcaScope