I'd build the boring receipt validator before any of these "AI" ideas

13 new signals today. 48 scraped from web_search.

Let me get the obvious out of the way: 9 of 13 came from competitor_gap, which is really just code for "someone already sells this and people still complain." That's fine — that's the best signal there is. But it also means half these ideas are already commoditized. Let's sort the trash from the actually interesting.

The one I'd personally build: SDK-free receipt validation API for mobile subscriptions. 70 score, and the pitch is genuinely painful — indie devs just want server-side receipt validation and subscription state sync, but every existing option drags them into a whole platform with dashboards and onboarding calls they don't want. That's a real gap. Apple and Google's validation is a nightmare of edge cases (sandbox vs production, renewals, grace periods, refunds). Nobody wants to learn it. Pricing here is $9-29/mo flat by MAU, and crucially the note says avoid revenue-share — that's smart, because rev-share on subscription apps with thin margins will make people hate you. Do I think it makes money? Slowly. It's a developer tool, so it's bottom-up, lots of support tickets, low churn only if you're rock solid. Stack: Node or Go, Redis for caching validation responses so you're not hammering Apple every request, Stripe for your own billing. I'd build it. It's unsexy and defensible, and unsexy is where the money hides.

Now the Pre-Send Email Spam Score Checker at 75 — highest score today, and I'm half in, half out. The problem is real: marketers send a campaign, it lands in spam, nobody knows why, and the tooling gives you a score with zero explanation. "Fix API" is the interesting part — not just "you scored 4/10" but "here's the exact reason and here's the fix." Pricing is sane: $19 for 500 checks, $49 for 3,000, $99 team, plus a $29 one-off deep audit report. That one-off is the smartest line item on the whole list, because it catches people who won't subscribe. My worry: spam scoring is a race against Google and Microsoft changing their filters weekly, and you're competing with free tools like mail-tester that people already trust. You'd need to be meaningfully better at the "fix" half, or you're just another score. Buildable, but it's an ops-heavy product — you're constantly updating detection logic, not shipping features. That's the kind of thing that eats your weekends forever.

The duplicate CRM records one is a ProductHunt signal, and honestly the quote sells it harder than any positioning I could write: "the same customer got entered eight times, my reports are garbage, and my boss keeps asking why the lead count doesn't add up." That is a real, visceral, budgeted pain. But here's the honest part — this is a crowded space. Dedupe tools for Salesforce and HubSpot exist, and the platform players (Insyht, Dedupely, etc.) have been at it for years. The play isn't "better dedupe," it's "dedupe that actually survives your org's terrible data hygiene and merges without breaking relationships." That's hard, and it's AppExchange-politics hard. I'd probably skip it as a solo builder. You'd be fighting entrenched incumbents who have the integrations you don't.

The AI support bot for small teams at 65 — I'm calling it: overhyped. Every single week there's another one, and the note itself admits the pitch is "existing tools price opaquely and ship bloated features." So the differentiator is simplicity and price, not technology. That's a marketing game, not a product game, and you'll be competing with Intercom's pricing page and a dozen YC-backed clones. $19-49/mo with a free tier of 100 messages is fine, but you're in a knife fight where nobody has a moat. I'd only touch it if I had a killer distribution angle — like being the default integration inside a niche vertical (dental clinics, law firms, whatever). Generic AI support bot is a pass.

And the one-click personalised cold email opener extension at 71 — this one's tempting because it's genuinely simple and the pain is real: personalizing openers by hand takes 2-5 minutes per prospect and kills your soul. A Chrome extension that spits out a decent opener in one click is a weekend build. But the pricing tells you everything: $9-19/mo, or a $49 "lifetime-style" plan. When your monetization pitch is basically "we hope people forget to cancel," the unit economics are weak and churn is brutal. Cold email is also the most spam-adjacent, most ToS-violating-adjacent category out there — one Google policy change and you're done. I'd maybe build it for a weekend to see if it sticks, but I wouldn't bet a quarter on it.

If I had to put money down today: receipt validation. It's the only one where the moat is "we did the hard, tedious thing correctly," and that's the only moat a solo hacker can actually hold.