4 X demand signals: August 1-8
Four real X requests survived the week's screen, from an AI-free visual journal to a personalized running-audio companion, each narrowed to a testable indie experiment.
Four requests made the cut
The clearest request this week is a date-first visual journal that does not use AI. A browser-based feed for useful targeted ads is more interesting as a constrained platform experiment than as a clean SaaS. Two lower-signal asks - a widget that prices your time and an adaptive audio companion for runs - need direct user validation before code.
Coverage: August 1, 08:00 through August 8, 08:00 in the channel time zone (UTC-05:00). Engagement means likes + reposts + replies; quote posts are excluded.
The ranking weighs raw engagement, author context, pain-point specificity, replies that name an existing solution, and whether a solo developer can test the smallest useful version. Engagement is a demand clue, not proof of a market. The repeated employee-exit request was left out because the same job appeared in the previous issue.
| Rank | Opportunity | Counted engagement | Author context | Verdict | Main constraint |
|---|---|---|---|---|---|
| 1 | Date-first, AI-free visual journal | 27 1 | 1,589 followers; verified account | Conditional build | Pinterest adjacency and active builders in the replies |
| 2 | One feed for useful targeted ads | 7 2 | 409 followers; verified account; former founder and software engineer | Conditional browser experiment | Ad capture, consent, and platform access |
| 3 | A widget that shows what your time costs | 3 3 | 172 followers; verified account | Weak signal, fast test | The hourly-value formula may feel arbitrary |
| 4 | Personalized audio for daily runs | 0 4 | 1,104 followers; verified account; CTO | Weak signal, concierge first | Content rights, freshness, and repeat use |
1. Date-first, AI-free visual journal
The demand signal: @missbananasart asked for something like Pinterest without AI integration. The requested workflow is specific: save inspiration every day, tag boards, browse by date rather than only by board name, and optionally view other people's date-based boards. The post appeared August 6 at 3:37 a.m. ET and drew 21 likes, one quote, and six replies. 1
This is not a clean empty market. Replies included two people offering to build it, and another person said they were already building something close. That is useful evidence of product pull, but it also means a new entrant needs a sharper wedge than "Pinterest, but smaller." 1
MVP: Make the date the primary index. A local-first browser extension or mobile app should save an image or URL, attach the capture date, add tags, and open a calendar view of everything saved that day. Add public boards only after the private archive works. Leave recommendations, generated tags, and social ranking out of version one.
Distribution path: Start with visual designers and illustrators who already collect references. Onboard 20 people manually, import a small sample of their existing saves, and watch whether they return to a date view rather than the ordinary board view. The active builders in the thread make the comparison product quality, speed, and privacy - not discovery.
Main risk: "No AI" is a product boundary, not a moat. The product also depends on browser capture and image hosting, while an active builder may reach the same workflow first. Charge only after users have retrieved old references repeatedly; a one-time burst of curiosity is not retention.
2. One feed for useful targeted ads
The demand signal: @rasha_hantash wanted an app that routes the targeted ads they already see into one scrollable feed. The reason is unusually clear: they do not want to block every ad because some targeted ads are useful. The post appeared August 7 at 4:50 p.m. ET and had seven likes, with no reposts or replies. The author has 409 followers, a verification mark, and describes themselves as a former founder and software engineer. 2
The tension is the product: preserve useful discovery while removing the interruption of the original feed. But this is one person's request, not a confirmed cluster. The technical and policy surface is also the business risk. A solo developer should not promise to collect ads from every network before proving that one supported surface can be captured with clear user consent.
MVP: Build a browser extension for one site. Let a user save an ad they are currently shown into a private shelf, preserve the destination and a small amount of context, and scroll that shelf later. Do not start by inferring why the ad was targeted or by copying ad inventories across networks. The first test should answer whether people revisit saved ads and click through, not whether the feed looks busy.
Distribution path: Recruit people who save ads for shopping, travel, or creative research. Give them a seven-day private shelf and ask which saved ads they revisit. If the product cannot retain the destination, explain the source, and honor deletion requests, it is not ready for a broader audience.
Main risk: The extension may be blocked by page changes or network rules, and the ads may carry sensitive targeting signals. The product would need a strict data-retention boundary and a supported-surface list. This is a platform-dependent experiment, not a general ad marketplace.
3. A widget that shows what your time costs
The demand signal: @KumarManik12 asked for an app or widget that calculates how much money their time costs while doing an activity for a given duration. The post appeared August 7 at 7:13 a.m. ET and had two likes and one reply. The account is verified, has 172 followers, and describes its author as working across technology, startups, fitness, and content. 3
The only reply asked whether the idea came from Naval Ravikant's "Aspirational Hour". The author said it was not. No reply named a complete product that matches the requested phone-widget workflow. 3 That is a gap in the thread, not proof that the category is empty.
MVP: Let the user set one hourly value, start a timer from a home-screen widget, and show the running cost. Add a calendar of sessions and a weekly total only after the basic interaction feels useful. Keep all activity names and amounts on the device; automatic activity surveillance would add anxiety without validating the core job.
Distribution path: Test with freelancers, consultants, and solo founders who already price work by the hour. A landing page and a working widget are enough for the first ten users. Ask them to use it for a week, then check whether they open the history, change the hourly baseline, or would pay for a persistent widget.
Main risk: The number is a chosen proxy, not an objective cost. It may motivate one person and annoy another. If users only enjoy the first-day novelty, the idea is a calculator with a notification rather than a durable product.
4. Personalized audio for daily runs
The demand signal: @marccampbell wanted an app that generates audio for a run based on his preferences: sometimes technology news, sometimes a lesson, sometimes a documentary, and sometimes a short science-fiction story. The post appeared August 4 at 1:00 p.m. ET and received no likes, reposts, or replies. The author is a verified CTO with 1,104 followers. 4
The request has a clear use moment but weak public validation. It also bundles several content products into one sentence. A solo developer should test the running habit first and postpone the full library of formats.
MVP: Offer a 20-minute, user-selected audio brief for one run type. A concierge workflow can assemble the source material, create the narration, and deliver one file before building a recommendation engine. Let the runner choose "news," "learn," or "story"; do not mix all three until one mode earns a repeat session.
Distribution path: Find ten runners through one local club or running community and deliver the first episodes manually. Measure completion, repeat use the following week, and which format people request again. Those observations matter more than a polished voice demo.
Main risk: Current-news audio brings freshness and rights questions; generated stories bring a different content-quality problem. Running also limits attention, so the product needs a reliable pause and resume experience. With zero engagement on the original post, this belongs in concierge validation, not a full build.
What did not make the cut
- Pantry-to-dinner planning: The post drew 42 likes, three reposts, and 16 replies, but replies named Supercook, other recipe apps, and using ChatGPT with photos of the pantry. The raw number is strong; the open gap is not. 5
- Searchable photo archives across events: The request was concrete, but replies named Memzo, Kwikpic, PhotoShelter, and Bynder as existing ways to tag and retrieve people or company images. That makes this an incumbent comparison problem, not an unserved request. 6
- A cross-tool AI usage audit: Replies pointed to
ccusage,better-ccusage, and Codex Bar; the original poster said one of the suggested tools worked perfectly. The thread itself closes the gap. 7 - Song identification from an Instagram link: A reply immediately named Shazam, and the author acknowledged the product while arguing that people need better awareness of it. That is distribution friction, not a new product category. 8
- Browsing an X account's posts by date: A reply named the TimeSkip Chrome extension and X's native date-filter search. The request is real, but the direct solution is already in the thread. 9
- BIP110 signaling tracker: The post drew 11 likes and nine replies, but several replies supplied existing statistics and tracker links. It is a solved information-discovery request for this window, not a clean build lead. 10
Validate before building
- Visual journal: Onboard 20 creatives, let them import a small archive, and record how often the date view is used. The product earns a second week only if it beats ordinary folders for finding old references.
- Targeted-ad feed: Support one web surface and test saving, revisiting, explaining, and deleting an ad. Stop if the capture cannot be reliable without over-collecting personal data.
- Time-cost widget: Ship the smallest widget to ten people for seven days. Ask whether they changed the hourly baseline, opened history, and would keep it installed.
- Running audio: Produce ten episodes manually, one format at a time. Keep the idea only if runners finish them and request another the following week.
The first signal has the cleanest workflow, but it also has visible competition. The other three are useful precisely because the evidence is thinner: each can be tested with a small audience before a founder commits to infrastructure, content costs, or a platform dependency.
References
- 1request on Xx.com
- 2request on Xx.com
- 3request on Xx.com
- 4request on Xx.com
- 5
- 6
- 7
- 8
- 9
- 10
Twitter 'I want an app that...' Demand Radar
Scrape 'I want an app that...' / 'someone should build...' posts on X, ranked by signal strength of real demand, surfacing project ideas for indie developers
This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.
Related content
- Sign in to comment.