Cosmos is glitchy, flashy, and filter-leaky on Google Play

Cosmos is glitchy, flashy, and filter-leaky

In 16 weeks you can make $1,200/month by building a better accessible human-only moodboard saver.

2/5
Opportunity score5 = build it
πŸ“ˆ3/5Market gap5 = large gap
πŸ”¨16weeksBuild timesolo, part-time
πŸ’°$1,200/moExpected MRRat month 12
4.7β˜… lifetime4.2β˜… now1M+ installs

Abstract

Cosmos has real pull as a Pinterest alternative, but the complaints point to a narrow solo-builder wedge β€” reliable saving, safer onboarding, stronger AI/NSFW controls, and accessibility β€” not a realistic chance to clone its full social discovery network.

Background

Cosmos: Search & Discover is a free visual discovery app for saving ideas, searching by color, keyword, or image, importing from anywhere, following tastemakers, collaborating on collections, and hiding AI content. It has 1,036,170 installs and a 4.7β˜… lifetime rating, so this is not a dead app β€” it is a popular app with visible execution cracks.

What it does

A Pinterest-like inspiration space for moodboards, art references, collections, search, import, and social discovery.

What changed

Recent reviews complain about slow loading, broken signup/login, haptic and layout regressions, flashing startup imagery, and filters that miss AI, porn, gore, or unwanted topics.

Why it is soft

A solo builder cannot out-network Cosmos, but can target users who want a calmer, safer, more reliable private moodboard tool rather than a broad social feed.

Strengths

Cosmos still has strong user love: reviewers praise the clean UI, artist credits, dark mode, import-by-link, lack of ads, search by color and hex, and the promise of filtering AI content. Several users explicitly say it feels like old Pinterest or that they replaced Pinterest with it, which means the habit and demand are real.

Clear positioning

The anti-ad, anti-AI, inspiration-first pitch resonates with artists and creative users who are tired of Pinterest’s current feel.

Useful discovery features

Color search, image search, collections, import, credits, and sharing are not gimmicks; users mention them as reasons they want Cosmos to work.

Large trust base

The 4.7β˜… lifetime rating and 1,036,170 installs make it hard for a new entrant to look more trustworthy on day one.

Market gap

The complaints cluster around reliability first, then safety/control, then onboarding and migration. The opportunity is not β€œbuild another Cosmos”; it is β€œbuild the dependable, private, accessible subset Cosmos users expected.”

Slow, broken, unstable core flows

The most common complaint theme is that loading, saving, scrolling, account creation, login, or collection creation fails at the exact moment users try to build a habit.

AI, porn, gore, and feed control

Users like that AI and suggestive filters exist, but several say unwanted AI, porn, gore, or repeated categories still get through despite controls.

Accessibility and hostile UX

The flashing startup sequence, haptic feedback, hidden buttons, tablet layout issues, and confusing posting flow create a wedge for an app that is calm by default.

Switching from Pinterest

A few users want Pinterest transfer or migration, and one complains that promised transfer was not possible in the app or web.

Can we have a flash warning at least when we open up the app and are immediately bombarded with 500 pictures flashing really fast as the opening?? I opened it last night for the first time in a while and it really hurt my eyes, I can't imagine what it would be like for someone with epilepsy. It's not a good idea.
β˜…β˜…β˜…β˜†β˜†Cosmos: Search & Discover Β· 2026-05-06 Β· 20 found this helpful
ive been using it for about a day now, so far i have only a few complaints. 1: the home page algorithm sucks, I spent like an hour just sitting refreshing the home page, saving things i actually want to see, and then nothing changes, its just the same stuff over and over, id like to see stuff i want to see, not borderline 🌽 in most cases. 2: it takes a bit to load, its not horrible but it can be annoying. It has potential, but right now, its not that good
β˜…β˜…β˜†β˜†β˜†Cosmos: Search & Discover Β· 2026-02-27 Β· 15 found this helpful
This app is amazing!! As an avid Pinterest user, I absolutely love to see the anti AI and ads!! It's a huge ask, but I would be so happy to see a feature that allows you to somehow transfer things from Pinterest. Like I said, I use Pinterest a lot and have tons of boards and thousands of pins saved, and it sucks to start over. Still, great app!
β˜…β˜…β˜…β˜…β˜…Cosmos: Search & Discover Β· 2026-06-16 Β· 12 found this helpful
  1. Do not build the whole social feed. Ship private moodboards, fast saving, import-by-link, collections, and search first; social discovery is where Cosmos has the strongest moat.
  2. Make accessibility a product promise. No flashing launch animation, optional haptics, obvious buttons, tablet-safe layouts, and a conservative content-filter mode directly answer repeated UX complaints.
  3. Win switchers with migration-lite tools. Offer bulk URL import, browser/share-sheet capture, and simple board organization before attempting a full Pinterest-transfer system.

Build complexity

Estimated, not derived: one developer, part-time, no funding, and no paid acquisition. This assumes a narrow Android-first v1 with web clip/import, private collections, basic search, accessibility-first UI, and moderation controls β€” not a full Cosmos/Pinterest clone.

WorkstreamWeeks
Android app shell, auth-light onboarding, local-first collections3.0
Save/import from share sheet, URL metadata, image upload3.0
Search, tags, color labels, board organization2.5
Accessibility-safe UI, tablet layouts, no-flash flows2.0
Basic AI/NSFW filtering controls and reporting2.5
Sync, backup/export, billing, Play Store launch3.0
Total16.0

Expected revenue

Estimated, not derived: with no paid acquisition, assume slow organic discovery from creators searching for a calmer Pinterest alternative. The model is freemium with a $4/month Pro plan for larger libraries, sync, backup/export, and advanced filtering.

Month 12, monthly
Installs3,000
Activated1,20040%
Retained60020%
Paid30010%
Revenue$1,200$4/month

Biggest challenges

The hard part is not making boards; it is getting trusted enough to replace a beautiful, already-loved app while avoiding the moderation, copyright, discovery, and migration traps that make visual platforms expensive to run.

Cold-start content

Without a social graph and tastemaker feed, a new app must feel useful as a private saver before it has community discovery.

No clean Pinterest migration

Users want thousands of pins moved, but a solo builder should expect brittle import paths and avoid promising full transfer unless it is actually reliable.

Moderation risk

AI, porn, gore, explicit results, and age gates are product promises that can trigger user distrust and Play Store risk if handled casually.

AI angle

There is a genuine AI angle, but it should be defensive rather than flashy: AI can classify saved images for likely AI-generated content, NSFW, gore, dominant colors, and tags, making the filtering and organization wedge cheaper than hand-labeling everything. It does not remove the need for user controls, mistakes handling, or moderation policy.

Where AI helps

Auto-tags, color extraction, visual similarity, and conservative safety labels can make a private moodboard feel organized without a large content team.

Where AI does not solve it

Users already complain that filters miss things, so the product must expose manual hide/report controls and never promise perfect detection.

Conclusion

Skip the full competitor; build the wedge only. Cosmos is too loved and too networked for a solo part-time clone to beat, but a smaller accessible human-only moodboard saver could attract users who mainly want reliable saving, calm UX, and stronger controls.

16 weeks

Time to v1

$4/month

Freemium Pro model

$1,200

Revenue at month 12

4.7β˜…

Fact that limits the opportunity

Get new reports by email

One email when a new opportunity report goes up. No spam, unsubscribe anytime.