Open dataset · Edition 1
The AI Visibility Index for Astrology Apps
Which astrology apps do ChatGPT, Perplexity and Gemini actually recommend — and what do they cite when they do? 50 dated audit runs. 65 apps. 219 source citations across 78 domains. Published in full, method included, so anyone can re-run it.
Vandro Labs is a Generative Engine Optimization agency for astrology and spiritual wellness apps — we work on getting apps named inside ChatGPT, Perplexity and Gemini answers. We built this index because we needed it for client work, then published the whole thing. What we do · Free audit of your app
Read this before quoting a number. Generative engines change their answers week to week. Every figure here is tied to a run date between 26 and 31 July 2026. If you are reading this more than about sixty days later, treat it as history, not as current state — that is exactly why we date everything and re-run on a schedule.
Which apps get recommended
Across 50 dated runs in July 2026, CHANI, The Pattern and TimePassages were each recommended in 22 runs, and Co-Star in 19. No app was recommended in every run, on any engine.
The column that matters is not the total — it is how far apart the three engines are. Read across the rows.
| App | All runs | ChatGPT n=17 | Perplexity n=16 | Gemini n=17 |
|---|---|---|---|---|
| CHANI | 22 / 50 | 6 | 7 | 9 |
| The Pattern | 22 / 50 | 10 | 5 | 7 |
| TimePassages | 22 / 50 | 8 | 5 | 9 |
| Co-Star | 19 / 50 | 7 | 4 | 8 |
| Astro Gold | 8 / 50 | 4 | 3 | 1 |
| Sanctuary | 6 / 50 | 0 | 0 | 6 |
| Solar Fire | 6 / 50 | 2 | 2 | 2 |
| Astro.com | 5 / 50 | 2 | 3 | 0 |
| AstroMatrix | 5 / 50 | 2 | 2 | 1 |
| Astro-Seek | 5 / 50 | 4 | 0 | 1 |
| Nebula | 4 / 50 | 4 | 0 | 0 |
| iPhemeris | 4 / 50 | 1 | 1 | 2 |
| Horos | 3 / 50 | 0 | 3 | 0 |
| Time Nomad | 3 / 50 | 0 | 1 | 2 |
| CUE Astrology | 3 / 50 | 1 | 1 | 1 |
| Staia | 3 / 50 | 1 | 1 | 1 |
Finding 1 — An app can lead one engine and not exist on the other two
Sanctuary was recommended in 6 of 17 Gemini runs and in zero ChatGPT and zero Perplexity runs. Nebula was recommended in 4 of 17 ChatGPT runs and in none on the other two engines.
These are not rounding errors. They are two apps with real market presence that are, functionally, invisible on two thirds of the surfaces where the question gets asked. Astro-Seek shows the same shape in reverse: 4 of 17 on ChatGPT, zero on Perplexity.
The practical consequence for an app team: “are we visible in AI?” is three separate questions with three separate answers, and a dashboard that averages them hides the only thing you could have acted on.
Finding 2 — Visibility is a frequency, not a state
We re-ran identical questions on different days. An app cited by an engine on one day was absent from the same engine on the same question two days later. Perplexity was the most stable of the three engines; Gemini the most volatile.
In one control run, a small app appeared in a Gemini answer and then vanished from a re-run of the identical prompt. Nothing about the app had changed. This is why every number in this index is a rate out of n runs rather than a checkmark, and why we treat any single screenshot — ours or a competitor's — as anecdote.
Finding 3 — The most cited source is not journalism
Of 219 citations recorded, 20% pointed to apps ranking themselves on their own comparison pages. Genuine editorial publications accounted for 5%.
| Cited source type | Citations | Share |
|---|---|---|
| App Store / Play Store listing | 59 | 26.9% |
| App ranking itself on its own comparison page | 44 | 20.1% |
| Directory, aggregator or ranking site | 30 | 13.7% |
| Another app blogging about the category | 26 | 11.9% |
| Official site of an app or of the niche | 15 | 6.8% |
| Genuine editorial publication | 11 | 5.0% |
| Reddit or forum thread | 8 | 3.7% |
| Dev-agency lead-generation blog | 7 | 3.2% |
| YouTube reviewer | 7 | 3.2% |
| Other or unclassified | 12 | 5.5% |
| Total citations recorded | 219 | |
Three consequences follow, and they are uncomfortable ones:
- Your store listing is a citation surface. It is the single most cited source type in this niche. We watched all three engines quote store copy — and in one case quote a developer's public reply to a user review — straight into an answer.
- Self-published comparison pages work. A pre-launch product with no app shipped and a waiting list in the hundreds was recommended alongside a decades-old desktop incumbent, on the strength of one honest comparison page.
- Bias is not being filtered out. Pages that rank themselves first with no disclosure are cited routinely. We had assumed the opposite and the data corrected us. We still build honest comparisons for clients — because they survive contact with competitors, audits and updates, not because an engine rewards virtue. Any agency telling you “you have to be honest or you won't get cited” has not measured it.
Finding 4 — Review count predicts almost nothing
In this niche, the number of App Store reviews predicts neither whether an engine recommends an app nor how much money the app makes.
Two of the apps in the leaderboard above sit outside the top twenty by review count. Meanwhile the app with the largest review count we sampled — over 42,000 reviews — was recommended in a small minority of runs. We had been using review count to prioritise which apps to study. It was the wrong sort order, and we published this finding rather than quietly fixing it.
How to read this index if you run an app
- Check all three engines separately. Your average is not a number that exists.
- Ask the question your user asks, not the one you would ask. “Best astrology app for beginners” and “professional astrology software for chart work” return almost disjoint sets of apps.
- Run it more than once, on more than one day, and record the date. One run is a coin flip.
- Look at what got cited, not just who got named. The citation list is the map of what you would have to be part of.
The full method, including the exact question set and session setup →
Want your app measured against this index?
We will run your app through the same question set on all three engines and send you the dated results — the recommendation rate, the questions where you lose, and who is cited instead of you. Free, no call required, reply by email.
Citing this index
Free to quote and republish with attribution (CC BY 4.0). Suggested citation:
Vandro Labs (2026). The AI Visibility Index for Astrology Apps, Edition 1, July 2026. https://vandrolabs.com/ai-visibility-index/