Ideas Radar: 2026-09-23

Three independent posts in three cities — Toronto, Thessaloniki and a California campus — all ask for the same missing product: a transit app that actually ingests service changes. The complaint is identical each time. You plan a trip in a mainstream maps app, arrive at the stop, and find a printed sheet taped to a pole announcing a diversion the app never mentioned. The data exists; agencies publish alerts, they just live in a feed nobody merges with routing. A transit app whose headline feature is "we route around today's closures, not last quarter's timetable" is a narrow product with an obvious daily use, and the fact that three unrelated people described it in one window is the strongest demand signal on this list.
Source: Reddit

A retail operator wrote what is effectively a finished product spec and framed it precisely: he is not trying to replace the point-of-sale system, he is trying to fix everything that happens between the POS and actually running the store. Walk the floor, scan a barcode, tap Low or Out, done; at ordering time the app assembles the week's marks into an order sheet and remembers what happened last time. Counting works the same way — scan, enter floor count, enter back-stock count, record what actually arrived against what was ordered. The clever part is onboarding: photograph a shelf section, draw lines separating products, and let the system read the packaging for names and retail prices instead of typing a catalogue in by hand. Add staged permissions for staff, managers and multi-store owners, plus captured customer requests for products you do not carry, and this is a real business.
Source: Reddit

Costco deliberately does not put item signs on its aisles, and someone wants to fix that from the outside: a crowdsourced, searchable map of every store showing which aisle and bay holds which item, with users reporting entries as out of date when things move. The elegance is that the data problem and the use case are the same action — people who cannot find something are exactly the people standing in the aisle able to record where it actually is. This generalises well past one warehouse chain to any large-format retailer that treats the treasure hunt as a feature, and the out-of-date reporting mechanism is the part most crowdsourced-map attempts get wrong.
Source: Reddit

A coach describes the cleanest example of a gap created by an acquisition rather than by neglect. He used an app called Technique to film athletes on a tripod, scrub the slow-motion footage locally, and drop timing markers at the first and second cone to get a split. Then it became a cloud product and the simplicity died. He is not asking for AI, analysis or a social feed — he wants local recording, slow-motion scrubbing and two manual markers. He also notes the vendor was uninterested in the feedback. Whenever a simple local tool gets bought and turned into a subscription platform, it leaves behind a user base that would happily pay for the original, and those users are unusually easy to find because they all complain in the same places.
Source: Reddit

A first-time drone builder wants what PCPartPicker is for desktops: pick components, see whether they actually work together. Anyone who has built an FPV drone knows this is a real compatibility matrix — flight controller to ESC mounting patterns, voltage ranges, protocol support, frame and motor and propeller geometry, video transmitter and receiver pairing — and that the knowledge currently lives in forum folklore and YouTube. The same window produced a near-identical ask for assembled-PCB quoting, which suggests the general shape is undersupplied: hobbyist hardware has plenty of component catalogues and almost no compatibility engines.
Source: Reddit

An expectant father describes a paywalled product failing at the thing it is prescribed for. Fetal kick counting is clinical guidance, and the available apps either cost real money or only count kicks inside a timed ten-minute window rather than tracking total movement across a day. His question is the damning one: it seems important, why is it hard to find a good free one? A well-made, free, all-day kick tracker with a clean export for a midwife appointment is a small piece of software with a genuinely medical use, and the current state of the category looks like an opportunity rather than a market verdict.
Source: Reddit

Someone with OCD wants to brain-dump thoughts into an app and have it sort them into categories automatically, specifically so the result can be handed to a therapist. The reason is the whole insight: manually categorising your own intrusive thoughts is itself an unbearable task for the person who most needs it done. This is a case where automatic clustering is not a convenience feature but the entire product, because the manual version has a psychological cost that blocks adoption. It generalises to therapy prep more broadly — the gap between an unstructured week of feeling and the fifty minutes you get to describe it.
Source: Reddit

Someone helping a friend clear out his late mother's house found ten-plus twenty-seven-gallon totes of beads, sorted by colour, each set in its own small box, and cannot find any venue to convert that into money rather than a donation. He asks the right question: is there an Etsy-equivalent for excess craft supplies, and do people hold shows for beads the way they do for stamps and coins? Estate liquidation of specialist hobby inventory is a recurring, emotionally loaded, entirely unserved problem — the supply arrives in one enormous lot at the worst possible moment for the seller, and the buyers are scattered across hobby communities that have no shared marketplace.
Source: Reddit

A one-line post carries a surprisingly large product: is there an app that keeps checking I stay alive. The category exists in fragments — dead-man's-switch scripts, emergency contact features buried in phone settings — but nothing consumer-grade aimed at the actual population, which is people living alone who are quietly aware that nobody would notice for days. The design questions are genuinely interesting and nobody seems to be working them: how often to check, how to avoid becoming a nagging notification, who gets contacted, and how to fail safe rather than fail loud.
Source: Reddit

A fly fisher travelling to unfamiliar water wants something that sounds like it should exist and does not: a map of where you can legally park and wade a river, with public and private boundaries drawn on it. He is explicit that his worry is trespassing, not fish. Access rights, parking, and riparian ownership are public records in most jurisdictions and are almost never assembled into a usable layer, which leaves every visiting angler asking a forum. The same map serves kayakers, swimmers and birders, and the rights data is the moat rather than the app.
Source: Reddit

An electric vehicle owner posted the same question to two adjacent communities: which features are actually drawing power right now? He had this on a previous combustion car and cannot find it on the EV, which is the ironic part — the vehicle where range anxiety is real is the one that hides the breakdown. He wants to optimise, not just observe. Per-subsystem energy telemetry surfaced as a simple ranked list is a small feature that manufacturers apparently consider uninteresting, and the aftermarket has not filled it either.
Source: Reddit

Someone planning a trip to the Canadian Rockies wants a fall-colour progression tracker and names a precedent: Ontario's provincial parks publish one, and he used to rely on it. This is the cleanest possible product brief — the thing exists, it works, it is confined to one jurisdiction, and the demand is obviously wider. Leaf-peeping drives real travel spending on tight timing windows, and the data can come from a mix of webcams, ranger reports and crowdsourced photos. The hard part is not technology, it is being the first to assemble it for a region that lacks one.
Source: Reddit

Someone juggling several low-priority hobbies wants a timer that enforces a limit rather than a habit. He has found apps that track time well and apps that set goals, but the goal-setting ones treat everything as a streak to maintain. What he wants is the inverse: fifteen to thirty minutes on this today, then stop, with decent statistics afterwards. That distinction — a ceiling rather than a floor — is a real design gap in the productivity category, which is almost uniformly built around encouraging more of an activity rather than containing it.
Source: Reddit

Someone in Bangalore is tired of rebooking cleaners daily through a gig platform and wants a service providing trained household help on a standing monthly arrangement. His complaints are specific and operational: cancellations and inconsistent cleaning. The gig marketplaces optimised for instant matching, which is the wrong optimisation for a recurring domestic relationship where the value is that the same person learns your home. A product that brokers continuity rather than availability — same worker, guaranteed schedule, managed substitution — is a different business from the incumbents even though it serves the same transaction.
Source: Reddit

Two posts in unrelated fandoms want the same missing product: an authoritative tracker of what has and has not been released. One wants a site listing every unreleased and leaked track for a musician, noting that another artist's fanbase built exactly that. The other wants a list of action figures including renders and name-only reveals that have not shipped yet. Both are asking for a release-state database — announced, revealed, shipped, cancelled — which every large fandom maintains badly across wikis and forum threads. The fact that it recurs across music and collectibles in a single window suggests the template, not the topic, is the product.
Source: Reddit

A Mac user wants a desktop widget showing the currently playing Spotify track with an audio visualiser, and has already done the technical scoping himself: there is no official Mac widget, the track name is reachable via AppleScript, and the visualiser needs a loopback audio driver to capture output. He is asking whether anyone has combined those pieces so he does not have to. This is a small, well-defined build where the whole value is that someone has already fought the audio-routing permissions so the user does not have to.
Source: Reddit

A student building a project needs sim racing telemetry converted into the racing line as a series of coordinates exportable to CSV, and notes the alternative would be generating the track centre line. Racing sims expose telemetry, and enthusiast tools consume it for lap analysis, but almost none of them treat the driven line as data you can take away. Export is the feature. Locking analysis inside a viewer is the default across a lot of hobbyist telemetry software, and being the tool that opens the door is a cheap differentiator.
Source: Reddit

The same person posted to two adjacent communities looking for a way to strip the constant hum from a mirrorless camera's built-in microphone, and added a sharp observation about why searching failed: every result was an advertisement, so none of it was objective. The audio-repair request is modest and probably solvable, but the second half is the more interesting gap — the entire category of "which tool actually does this" has been captured by affiliate content, and the person searching knows it and has no alternative to asking strangers.
Source: Reddit

Someone bought a grab bag of mixed vintage toy parts at a church sale and wants to photograph the pile and be told what he has and what is missing to complete figures. This is a well-shaped computer vision problem with a passionate and specific user base: parts identification against a known catalogue, then completion-gap analysis against known assemblies. Collectors already maintain the reference data in wikis. Nobody has pointed a camera at it, and the same mechanism would work for construction sets, model kits and board game components.
Source: Reddit

A mature student trying to choose a postgraduate programme describes being unable to search the thing that actually matters. He knows roughly what he wants to research and has been told to look for doctoral funding that includes a master's, which means the real query is "which academics are working on this, and which of them have money attached" — and there is no way to ask that. Institution-first directories are the wrong shape for a research-first decision, and the underlying data, faculty research interests and funded positions, is public but scattered across thousands of department pages.
Source: Reddit

A beginner learning Norwegian wants a podcast structured like an actual lesson — introduce a grammatical concept, then give examples — rather than the two formats that dominate: full-speed native conversation, or slow storytelling with no instruction. The gap is real across most languages outside the big four, and the observation generalises: language audio content is bifurcated into immersion and entertainment, with the pedagogically structured middle largely unserved for anything smaller than Spanish.
Source: Reddit

A controller owner suspects his hardware is misbehaving rather than his own muscle memory, and wants a site or app to test it and find out. Stick drift, unintended inputs and phantom button presses are common enough to be a running joke, and the diagnostic — visualise every input, log anomalies over a session, produce something you could attach to a warranty claim — is straightforward. The framing matters: the user cannot tell whether the fault is the device or himself, and a tool that settles that question has clear value at the moment of purchase regret.
Source: Reddit

A cruise passenger discovered that the ship's photo gallery becomes inaccessible the moment you walk off the ship, and wanted to buy digital copies after the fact. Photos were taken, printed and sold, and the digital versions simply vanish behind a door that only opens on board. That is a deliberate scarcity tactic that leaves money on the table, and it describes a general pattern worth attacking: venue-captured media — cruises, theme parks, race events, ski resorts — where the photographs outlive the customer's access to them.
Source: Reddit

Someone wants per-app internet blocking on a normal phone, naming a hardened mobile OS as the thing that already does it properly. This is a permission that most users assume exists and does not: deciding that a specific installed app may never reach the network. Workarounds involve local VPN tricks, and the request keeps recurring because the mental model — apps ask for camera and location, why not network — is reasonable and the platforms have simply declined to expose it.
Source: Reddit

Two separate trades posts describe the same shape of missing tool, and it is the seventh consecutive window this pattern has appeared. One wants a way to remove a three-inch drain cap that will not loosen. The other is replacing hydraulic lifters in an engine and can extract the accessible ones, but the deep ones only rise a few millimetres and there is nothing to grip. Both are controlled extraction inside a confined bore where you cannot get a hand or a conventional puller in. The people posting are not asking for advice on whether to do the job — they have done most of it and hit a specific geometric constraint that no catalogue tool addresses.
Source: Reddit

Not a request but a demonstration, and it belongs here for what it says about who can now build hardware tooling. Someone reverse-engineered the serial shooting-data protocol of a discontinued film camera — command set, baud and framing, the multi-byte record layouts, a seven-hundred-entry conversion table, custom settings packed into four bytes — verified it byte-for-byte against two original Windows applications on real hardware, and shipped a GPL-licensed Mac application. Asked whether he used AI, he said yes, plainly, and named the coding agents. The interesting part is the category: obsolete-hardware protocol support is exactly the kind of work that was never worth a person's month and is now worth an evening.
Source: Reddit

A perfumer making small batches cannot find a reasonably priced supplier of bottles that ships within Canada, and the only answer circulating is to order from overseas marketplaces. Small-batch packaging sourcing is a recurring bottleneck for craft producers in markets that are too small for domestic distributors and too far for cheap freight, and the gap is distribution rather than manufacturing. A curated regional supply catalogue with real shipping costs attached would serve soap, candle, cosmetics and sauce makers equally.
Source: Reddit

Someone wants a search tool that returns actual resources rather than sponsored placements — his examples are free places to learn to code and open-access journal material — and explicitly rejects both mainstream search and a chatbot as substitutes. The complaint is worth taking seriously even though the ask is vague, because it names a specific failure: the commercial layer has grown thick enough that people now assume any recommendation they read has been paid for. A curated, transparently unsponsored directory for narrow high-intent categories is an old idea that the current state of search has made new again.
Source: Reddit