App Ideas That Don’t Exist Yet: 14 Real Ones, Checked Against What Already Exists

By Post your build ideas · 30 October 2026 · 21 min read

Truly new app ideas are rare. For every idea on this page we found something close in an app store, a browser store or on GitHub. What “doesn’t exist” usually means is one of three gaps: nobody has built it for this audience, at this price, or this simply. These are fourteen ideas posted on Post your build ideas, each with the closest existing product we found on 30 October 2026.

Why are most “app ideas that don’t exist” lists wrong?

Most lists promising app ideas that do not exist are written without searching for the idea. Someone writes “an app that tells you if the fruit is ripe”, publishes it, and a reader finds three apps for it in a minute. The list was not lying on purpose. It confused “I have not seen this” with “this is not there”. The App Store and Google Play hold millions of apps, and a simple idea has usually been tried at least once.

A more useful question is where does the existing product stop? An idea is worth building when something close exists but leaves a gap you can name. There are three gaps that come up again and again, and every idea below is tested against them.

Three honest meanings of “this doesn’t exist”
The gap	What it means	Example from the research below
Not for this audience	It exists, but for another country, platform or group	Pet feeding logs exist, but several lean on Apple’s iCloud, which leaves Android households out
Not at this price	It exists, but costs more than the job is worth, or is free where it should be paid	Neighbour tool lending exists as rental platforms; the idea is lending with no payments at all
Not this simply	It exists, but buried in a bigger product or behind a long setup	Appliance running costs are explained in guides, but we found no one-screen tool

Read the entries below with that table open in your head. If you can name the gap in one sentence, you have an idea. If you cannot, you have a category that is already served. The board that these ideas come from is public: browse everything on the idea board, or post your own if you have one you will never build.

How the ideas were picked: we chose fourteen live ideas across different categories, read each one in full on its own page, and then ran a quick search for the closest existing thing. We paraphrase what the poster wrote and add our own view of the hard part. Six of the fourteen were posted by the site’s operator or its own account to seed the board, and we label them. The poster’s pricing thoughts are described in words, not figures. No build is attached to any of them on the board as of 30 October 2026. That does not mean nobody has built something similar elsewhere, as the “closest thing” lines show.

Which everyday-life app ideas has nobody built well?

These seven solve a small, annoying moment: waiting on hold, reading a parking sign, choosing a game, standing in an airport. Each is easy to explain in one sentence, which is a good sign. Each also has a hard part that a quick prototype will not solve, which is the honest part.

01 Hold-for-you caller

An app that sits through the phone menu and the hold music for you, then rings your phone the moment a real person picks up.

Posted by: The site’s operator
Who it’s for: Anyone who keeps putting off a call to a bank, an airline or an insurer
First version: Cover the phone menus of the biggest few hundred companies, wait on hold, ring the user when a human answers
Hard part: Phone menus differ per company and change without notice; and the hand-off has to land at the right second
Poster’s pricing thought: A small fee per successful connection, or a subscription for unlimited use
Closest existing thing: Hold For Me, a paid AI assistant that waits on hold and calls you back

The poster’s own analysis is the most useful part: the hard problem is not machine learning but a maintained map of phone menus, kept current by users correcting wrong paths. The app must never speak for you. At most it tells the agent it is connecting the customer. The moment it negotiates, trust and liability are gone.

The closest thing we found on 30 October 2026 was Hold For Me. A directory listing describes it as limited to US phone numbers, which we did not verify beyond the listings. A news report also says Google was testing a hold feature. So “nobody has solved it” is too strong. The open gaps are other countries and the strict never-speaks rule. Read the original idea.

Read “Hold-for-you caller” on the board and copy its build prompt →

02 Parking sign reader

Point your phone at a pole of stacked signs and get one answer: you can park here until 6pm, or you cannot today.

Posted by: The site’s operator
Who it’s for: City drivers facing four signs on one pole
First version: One city with dense, confusing signage; photo in, one answer out, including “probably yes, but check”
Hard part: The rules behind the sign: public holidays, temporary event restrictions, what “except permit” means in that district
Poster’s pricing thought: Free for a few checks a month
Closest existing thing: ParkBuddy, an iOS app that reads a snapped sign and says yes or no

This is the most crowded idea on the page. A search turned up several photo-based sign readers, including ParkBuddy and the Parksy sign scanner, and the original “Can I Park Here” app dates back to around 2012. These are developer claims from store listings, and we did not test any of them.

What the poster adds is a stance: a confident wrong “yes” costs a tow, so the app must say when it is unsure. Whether any competitor handles holiday exceptions, we could not tell from the listings. If you build this one, your edge is local rule data and honest uncertainty, not the camera. Read the original idea.

Read “Parking sign reader” on the board and copy its build prompt →

03 Halal and kosher scanner

Scan a barcode and get one word, halal, kosher, neither or unclear, plus the ingredient that decided it.

Posted by: A named poster (peekly_app)
Who it’s for: Shoppers who follow halal or kosher rules and read labels in the aisle
First version: One country, one certification body, a verdict plus the deciding ingredient
Hard part: The data: certifiers differ by country and school of thought, and gelatine can go either way depending on its source
Poster’s pricing thought: Free to scan, because it stops being useful if it costs something in the aisle
Closest existing thing: HalalReady, one of many halal scanners

The poster’s key point is that “unclear” must be a first-class answer, or the app loses trust the first time it is wrong. That is a design decision, not a feature.

The closest thing we found on 30 October 2026 was a whole shelf of halal scanners. HalalWise and HalalReady are two, and their verdict labels differ (halal, haram, needs review, unknown), which suggests there is no shared standard. Our one search did not surface a kosher equivalent, and we did not check whether any of them names the ingredient that decided the verdict. The gap here is kosher, plus explanation. Read the original idea.

Read “Halal and kosher scanner” on the board and copy its build prompt →

04 Shared pet feeding log

A tiny app where anyone taps “fed” and everyone sees the last meal, treats and medicine, with a nudge if nobody logged breakfast.

Posted by: A named poster (Marta K.)
Who it’s for: Flatmates and families where several people feed the same cat or dog
First version: One tap to log a meal, a shared last-fed view, a push reminder if no meal is logged by a set time
Hard part: Getting every flatmate to install it and tap it, every time
Poster’s pricing thought: Free, maybe a small subscription for several pets
Closest existing thing: Fed?, an iOS feeding tracker that records who fed and when

The origin story is a real household: three people feed one cat and nobody knows if she has had breakfast. That is a good sign, because the poster is the user.

This one exists many times over. Fed? syncs through Apple’s iCloud and puts family sharing behind a paid tier, per its listing. Others include Pet Feeder App and a 2015 app called DogSync. A tool tied to iCloud may not suit households with Android phones, so a cross-platform version is a real gap. Whether any of them reminds you when nobody has fed the pet is something we did not check. Read the original idea.

Read “Shared pet feeding log” on the board and copy its build prompt →

05 Which friends own this game

Before you buy a co-op game, see which of your friends already own it.

Posted by: A named poster (Greta)
Who it’s for: Anyone who buys games to play with friends
First version: Link one platform account, show friends who own the game on a store page
Hard part: Friend lists and libraries sit behind each platform’s own login and privacy settings
Poster’s pricing thought: Free, with affiliate links to the stores
Closest existing thing: Steam itself, which forum posts say lists friends who own a game on its store page

The poster says no store page tells you this and that it decides the purchase more often than reviews do. That is half true. Forum posts describe a “friends who own this game” list on Steam store pages, and a 2015 reply says it can miss friends who never launched the game. We could not find a current Steam help page confirming the feature, so check a game page while signed in.

That makes the real idea narrower and better: one view across stores and consoles, not inside Steam. The hard part is that no single company holds all those friend lists. Read the original idea.

Read “Which friends own this game” on the board and copy its build prompt →

06 Layover verdict

Enter your connection and get one honest answer: yes with time to spare, yes but tight, or absolutely not.

Posted by: The site’s own account
Who it’s for: Travellers with a long connection who wonder whether to leave the airport
First version: Connection in, verdict out, accounting for immigration queues, transit time into the city and the airline’s own cutoff
Hard part: Immigration queue times and transit rules are not public live data, and refusing to hedge is risky when you are guessing
Poster’s pricing thought: A small one-time fee
Closest existing thing: Chasing Whereabouts layover planner

The poster’s point is that existing tools show the layover length and let you do the maths. The value is in refusing to hedge. We found several layover calculators that subtract a fixed overhead from the connection time, and a Kayak guide on whether to leave at all. We did not test whether any of them use airport-specific queue data.

Our view: the verdict is easy to build and hard to defend. Start with a handful of hub airports where you can source real numbers, and say “not enough data” elsewhere. Read the original idea.

Read “Layover verdict” on the board and copy its build prompt →

07 RSVP that confirms

People click “going” and then do not come. This asks again 48 hours before, in one tap, and shows the organiser two numbers: said yes, and confirmed yes.

Posted by: A named poster (Trend)
Who it’s for: Anyone who has ordered catering for forty and fed twenty-two
First version: Import a guest list, send a one-tap confirmation 48 hours before, show said-yes versus confirmed-yes
Hard part: Guests ignore a second ask, and events usually already live on another invitation platform
Poster’s pricing thought: Free up to a guest count, a one-time fee above it
Closest existing thing: Luma’s automated reminders, sent one day and one hour before

The two-number display is the whole idea, and it is a good one. An organiser makes decisions on the second number, not the first.

The closest we found: Luma sends fixed reminders at one day and one hour before an event, per its help page. The attendance-management tool Solidarity.tech has a confirm button in its day-before email. We did not find a consumer invitation tool with a 48-hour confirmation step, but one search is not proof. Read the original idea.

Read “RSVP that confirms” on the board and copy its build prompt →

Which money and home app ideas are still open?

Money and home ideas have a built-in advantage: the user can feel the saving. They also run into a trust problem, because people hand over photos of their flat and their bank history only to products they trust. Five ideas, five different answers to that.

01 Move-in day photo report

A guided walkthrough that says exactly what to photograph, timestamps everything, and exports one PDF that you and the landlord sign on the spot.

Posted by: The site’s own account
Who it’s for: Renters who might lose part of a deposit over damage that was already there
First version: A checklist, a camera and a PDF: every wall, the inside of the oven, each window frame
Hard part: A timestamp shows when, not who; deposit rules differ by place
Poster’s pricing thought: A one-time fee per move
Closest existing thing: RoomProof, which photographs each room and exports a signed PDF

Several apps already do this, including RoomProof, Tenant Trace and RentProof. Landlords on a BiggerPockets thread argue that the signed baseline matters more than any app, and one wrote that a timestamp tells you when, not who. That is the real lesson for a builder: the signature step is the product.

So the idea exists, and the differences would be execution and who it speaks to. We could not tell from store listings which of these start from the tenant’s side. Read the original idea.

Read “Move-in day photo report” on the board and copy its build prompt →

02 Neighbourhood tool library

A hyperlocal map where neighbours list the tools they lend, request one for a date and hand it over at the door.

Posted by: A named poster (Sam O.)
Who it’s for: People who own a drill they use twice a year, and neighbours who need one
First version: A map, date requests and a simple “returned on time or not” reputation, no payments
Hard part: The cold start: a map with three tools on it is useless, so density per street is everything
Poster’s pricing thought: Free, sponsored by local hardware stores
Closest existing thing: Peerby and Lently

Neighbour lending apps exist. A search also found rental-style platforms, one subscription-based and one described as turning neighbours into contractors, and non-app tool libraries such as the Phinney Tool Library in Seattle. The no-payments design here is different from those. It is closer to lending a ladder over a fence.

Our view: do not build a national app. Pick one block or one housing estate, seed it with ten tools yourself, and only then think about a map. Read the original idea.

Read “Neighbourhood tool library” on the board and copy its build prompt →

03 Was this actually on sale?

A browser extension that shows twelve months of price history for the thing in your cart, because half of “50% off” was that price three weeks ago.

Posted by: A named poster (Alba)
Who it’s for: Anyone shopping during a sale event
First version: An extension that reads the product on the page and plots its past price
Hard part: Price history for an arbitrary shop has to come from somewhere; most existing history covers Amazon
Poster’s pricing thought: Free
Closest existing thing: CamelCamelCamel and Keepa for Amazon; Lowvyn for a Firefox add-on that compares against the average price

Price trackers are a mature category. A Reddit user used one to show that an Amazon seller’s Black Friday “discount” was a return to the price it had held for months. What our search did not turn up was a tool that checks the cart, rather than one product page at a time.

Our view: the price data is the business. The extension is the easy half. If you can source history from shops other than Amazon, you have something. If you cannot, you are the fifteenth extension. Read the original idea.

Read “Was this actually on sale?” on the board and copy its build prompt →

04 Subscription graveyard

Upload a bank export, find every recurring charge, see which ones you have not “used” lately by your own check-ins, and open the cancellation page in one tap.

Posted by: A named poster (Mo)
Who it’s for: People paying for things they forgot about
First version: CSV in, list of recurring charges out, a check-in per service, a link to each cancellation page
Hard part: Matching messy merchant names, keeping cancellation links alive, and getting people to check in
Poster’s pricing thought: A one-time fee
Closest existing thing: ChargeSift, a browser extension that finds recurring charges from bank exports on your device

The no-bank-login approach is already taken. An open-source Ruby gem, find-subscriptions, scans CSV exports. SubFind on the App Store does the same on iPhone. Some tools add cancellation guides too.

The part we did not find is the check-in: asking you whether you used something, instead of assuming a charge is waste. That is a small, cheap difference, and it is the one the poster cares about. Read the original idea.

Read “Subscription graveyard” on the board and copy its build prompt →

05 Appliance running cost

Photograph the energy label or type the model, add your electricity price, and see the yearly running cost next to the purchase price.

Posted by: The site’s own account
Who it’s for: Anyone about to buy a dryer, fridge or washing machine
First version: Model or label in, price per kWh in, a two-column cost comparison out
Hard part: Usage: published running costs differ mostly because they assume different cycles per year
Poster’s pricing thought: Free; how it gets funded is left as an open question
Closest existing thing: A formula in consumer guides such as Which?, not a dedicated tool

This is the one idea here where the closest thing we found was articles, not an app. The method is public: kWh per cycle, times cycles per year, times your tariff, added to the purchase price. Which? frames it as purchase price plus accumulated annual energy cost. A one-search check can miss a niche tool, so treat that as “not found”, not “not there”.

The poster calls it simple maths, badly presented. That is the “not this simply” gap in its purest form. Your risk is the funding question the poster flags: a free calculator needs a reason to exist beyond the build. Read the original idea.

Read “Appliance running cost” on the board and copy its build prompt →

Which developer tool ideas are still unbuilt?

Developer ideas are easier to test because you are your own first user, and the closest existing thing is a search on GitHub away. Two on the board show a pattern: a real workaround exists, and the idea is the polished version of it.

01 Closest-to-finished repo ranker

Point it at your GitHub and it ranks your abandoned repositories by how close each one is to shippable, not by how much code is in it.

Posted by: The site’s operator
Who it’s for: Anyone with a graveyard of half-built side projects
First version: Read commit history, open TODOs and empty tests, ask two or three questions per repo, output one sentence such as “about three evenings from launchable”
Hard part: “Close to done” is not visible in code alone: a project can be mostly built and still have no login
Poster’s pricing thought: Free for public repos
Closest existing thing: Project Graveyard, a Claude skill that scores local abandoned projects

The closest thing we found on 30 October 2026 scans local repositories and gives each a “resurrection” score. Another, Commitment Issues, turns a repo into a death certificate and can sort a profile into dead, struggling and alive. Neither, as far as the descriptions go, rates a repo by how far it is from a release. We did not run either one.

Because the poster admits the code cannot answer the question alone, the product is the two or three questions per repo. That is where a prototype should start. If you want to learn what a first version would involve, validate the idea first before you write the scanner. Read the original idea.

Read “Closest-to-finished repo ranker” on the board and copy its build prompt →

02 Env var diff between deploys

Diffs your environment across two deploys and shows which variables moved. Names only, never values.

Posted by: A named poster (Kalina)
Who it’s for: Developers debugging “it worked yesterday”
First version: Connect one hosting provider, pick two deploys, list added, removed and changed names
Hard part: Needing read access to each host, and proving it never touches values
Poster’s pricing thought: A monthly fee per project
Closest existing thing: A manual routine in Vercel’s docs; venv-sync for local-versus-project names

Vercel’s documentation describes pulling the variables for each environment and comparing the files yourself. The venv-sync package flags variables present in a local file but missing from a Vercel project, and masks values. Both are across environments, not between two deploys in time. In one search we found nothing that does the second, but we did not check every host.

The first thing to check before writing code is whether your host even records which variables a past deploy used. If it does not, the idea needs to start recording from the day it is installed. Read the original idea.

Read “Env var diff between deploys” on the board and copy its build prompt →

How many of these ideas already exist in some form?

All fourteen have a close relative. None came back as “nothing found”. The table is the honest summary of this page: it shows how a real idea looks when you check it, which is mostly crowded with one clear gap.

Fourteen ideas, the closest existing thing, and the gap
Idea	Closest existing thing	The gap
Hold-for-you caller	Hold For Me	Other countries; a strict never-speaks rule
Parking sign reader	ParkBuddy and others	Local rule data and honest uncertainty
Halal and kosher scanner	HalalReady, HalalWise	Kosher; naming the deciding ingredient
Shared pet feeding log	Fed?, Pet Feeder App	Android households; a “nobody fed her” reminder
Which friends own this game	Steam’s store page	A view across stores and consoles
Layover verdict	Layover calculators	Airport-specific queue data
RSVP that confirms	Luma reminders	A 48-hour confirmation and a two-number view
Move-in photo report	RoomProof, Tenant Trace	Tenant-first design and one-move pricing
Neighbourhood tool library	Peerby, Lently	No payments, one street at a time
Was this actually on sale?	CamelCamelCamel, Lowvyn	Cart-level checks beyond Amazon
Subscription graveyard	ChargeSift, SubFind	“Did you use it?” check-ins
Appliance running cost	Guides and formulas	A one-screen tool
Repo ranker	Project Graveyard	Rating by distance to release
Env var diff	Vercel docs, venv-sync	A diff between two deploys in time

Several of the gaps above are small. That is normal. A product does not need a whole new category; it needs one thing that the nearest alternative does not do. If you want the board’s wider picks, our list of the best app ideas of 2026 ranks ideas by how buildable they are.

How can I check in 15 minutes whether my app idea already exists?

You do not need a research budget. You need a job statement and six searches. The point is not to prove the idea is new. It is to find the three closest products and write down the one thing they do not do. Set a timer and stop at fifteen minutes.

Two minutes: write the job. One sentence, with no nouns from your solution: “know whether I can park here”, not “a parking sign OCR app”. People search for the job.
Three minutes: App Store and Google Play. Search the job phrase and three synonyms. Open the top results and read the one-star and two-star reviews. Those reviews are a free list of what the existing product leaves out.
Two minutes: Product Hunt. Search the same phrases. Products that launched and faded still tell you someone tried and what they called it.
Two minutes: GitHub. Search the job phrase. An open-source version tells you the core is buildable, and its open issues tell you what users ask for.
Three minutes: Google. Search “app” plus the job phrase, then “alternative to” plus the best name you found. The second search shows who competes with whom.
Two minutes: Reddit. Search “is there an app that” plus the job phrase. Threads show the names people suggest, and the replies that say “I wish” are the unmet demand.
One minute: write it down. The three closest products, and one sentence that begins “unlike all three, mine…”. If that sentence will not come, the idea is already served.

Search phrases to copy and fill in

app that [the job in plain words]
is there an app that [the job in plain words]
[the closest product] alternative
[the job in plain words] site:reddit.com
[the job in plain words] open source
[the job in plain words] extension

For a faster first pass, the free Idea Checker gives a quick second opinion. When the idea survives, the next step is testing demand before you build, which is covered in how to validate an app idea.

Does an existing competitor mean I should stop?

No, but it should change what you build. A competitor is evidence that people pay attention to the problem, which is the hardest thing to prove. What it should stop is the version of your plan that copies the competitor feature for feature.

Look at what actually separated the ideas above from what exists. The gaps were specific and checkable:

Platform. Some pet feeding logs sync through Apple’s iCloud, which leaves households with Android phones out.
Region. One hold-on-the-phone assistant is described as limited to US phone numbers.
Coverage. Halal scanners are plentiful and use different verdict labels; a kosher one did not turn up.
Pricing model. Neighbour tool lending exists as rental and subscription products. A free, no-payments version is a different thing.
Stance. A parking reader that says “probably yes, check the holiday rules” is a different product from one that says “yes”.

The other half of the answer is cost. If a competitor is free, funded and loved, think hard. If it is mediocre, expensive or abandoned, you have room. Either way, the next move is the same: pick a gap, build the smallest version that covers only that gap, and show it to the people who complained in those one-star reviews.

If you have an idea but not the time, you can post it on the board and someone else may build it. If you want to see where else to put it, read where to post app ideas. If you are the builder, browse the full idea board, copy the generated build prompt on any idea and paste it into Claude Code, Cursor or Lovable. If you ship one, the idea’s page has a “Built this?” link so you can attach the build.

Frequently asked questions

Are there any app ideas that truly don’t exist?

Very few, and almost none of the simple ones. Every idea we checked for this page had a close relative in an app store, a browser store or on GitHub. The useful question is not whether something exists, but who it leaves out: another country, another platform, a different price, or a simpler version. Those gaps are real, and they are where small builders win.

How do I find app ideas that people actually need?

Start with complaints, not brainstorming. Search Reddit for “is there an app that” and read one-star reviews of the nearest existing app. Both list what people want and cannot get. You can also browse the public idea board, where each idea says who has the problem. Then run the 15-minute existence check before you build.

Can I build one of the ideas on this page?

Yes. Every idea on the board is free for anyone to build, and each has a generated build prompt you can paste into Claude Code, Cursor or Lovable. If the poster left contact details, you can write to them, but you do not need permission. When you ship, use the “Built this?” link on the idea to attach your build.

What if my idea already exists?

Write down the three closest products and one thing none of them does. If you can finish the sentence “unlike all three, mine…” with something specific, a different audience, platform, price or level of simplicity, the idea is still worth building. If you cannot, pick a different gap or a different idea. Existing competitors prove demand; they do not close the door.

Is it safe to post an idea publicly?

Ideas are cheap and execution is hard, so most people who post are glad to see one built. Posting on Post your build ideas is free and needs no account or email. Publishing is permanent: you cannot edit or delete an idea afterwards, apart from removal of posts about a real person through the report page. Only post what you are happy to give away.

How do I post an app idea I’ll never build?

Open the submit page, type a name, a title of up to 90 characters and the idea in up to 4,000 characters. Pricing thoughts and a contact are optional. There is no account and no email. Your idea gets a generated build prompt and a page that builders can vote on. Describe the problem and who has it, not just the feature.

Created with AI assistance and edited for this site. Examples are illustrative and do not represent customer research. About the publisher.