I built a tool that finds niche communities where your target customers actually hang out (free demo)

Update for everyone who gave feedback here: NicheRadar is live on Product Hunt today → https://www.producthunt.com/posts/nicheradar-2

Several of your suggestions (confidence scores, B2B/B2C filtering, adjacent communities) are named in the launch as the roadmap. If you have a minute to leave honest feedback there, it would mean a lot.

@geolxl Quick follow-up — those three features you suggested shipped: confidence scores on every community, B2B/B2C audience filter, and adjacent pain-point communities. Your feedback went straight into the product. Appreciate you taking the time.

@LalitaRnD Your verification question stuck with me — replied above with the honest breakdown of what's real vs. estimated. Appreciate you pushing on that.

To both of you: we're starting to think through what a paid version of NicheRadar could look like, and the most useful thing right now isn't a survey — it's a short conversation with people who've actually used it. If you'd be open to sharing thoughts on what would make this genuinely valuable for your work, I'd love to connect. Drop your email here or reach me at [email protected] — happy to keep it off the thread.

Quick update for everyone following this thread — shipped a few things based on the feedback here:

What changed:

1. The results page now shows example output before you scan, so you know exactly what you're getting before you commit to trying it.

2. Community promo rules are now classified more accurately. Previously everything defaulted to "Strict" — that was a bug in the AI prompt, not the actual reality of these communities. Fixed.

3. The landing now auto-detects your browser language (EN/ES), so non-Spanish visitors no longer land on a Spanish page wondering what happened.

What's next on the roadmap:

The question @LalitaRnD asked — "how do you verify community activity, not just size?" — has been living rent-free in my head. The architecture for fetching real rules and recent post titles from Reddit is built. The blocker right now is that Reddit blocks cloud provider IPs for unauthenticated scraping. Working through it.

When that ships, the tool goes from "here are communities that probably discuss your problem" to "here's what people were literally posting about this week." That's the version worth paying for.

If you want early access to that version, the waitlist is at radar-v2-theta.vercel.app — free scan included, no card.

This is a very real problem. Finding companies is easier than finding the actual communities where people discuss the pain openly.
I’m building MeetIQ by Acjen AI, and one thing I’m learning is that “target customer” is too broad unless I can see the exact conversations: long recordings, meeting follow-ups, client notes, missed action items, etc.
I like that you include self-promotion rules and active pain points, because those are usually more useful than just community size.
Curious — how do you verify that the communities returned are actually active and not just theoretically relevant?

Interesting idea. As a senior ERP, full-stack, AI integration, and automation engineer, I've seen many technically strong products struggle not because of the product itself, but because founders couldn't identify where their target users actually spend time. Most growth tools optimize for company discovery, while community discovery remains largely manual and fragmented. I especially like the focus on pain points and founder-friendliness rather than just community size. I'd be curious to see how well NicheRadar performs for highly specialized B2B and enterprise niches where traditional acquisition channels are expensive and slow.

The self-promo and rule-check part is the hard bit, agreed. A raw list of subreddits or Discords is easy enough to generate, but founders usually stumble on the first post angle and get written off right away. With DictaFlow, we've found the useful filter is basically this: would the post still help the community if you deleted the product name? If you can match each community with that first useful angle, things get a lot more actionable.

Useful timing — we're building CurbBay, a marketplace for homeowners to rent unused garages/driveways. The hard part isn't listing automation, it's finding homeowners with latent inventory before they search for it.

I'd use this if it could surface local/homeowner communities (HOA boards, neighborhood forums, storage/parking groups) and include a suggested first non-spammy post angle. Happy to be a test case if you want one.

The self-promotion rules part is probably the killer feature here. A list of communities is easy to generate; knowing whether you can actually show up without getting banned is what saves time. I’d maybe add a “first useful post idea” for each community.

The distinction between finding companies and finding communities is what caught my attention. A lot of early-stage founders know who they want to sell to but not where those people already gather. If the recommendations consistently surface places with active conversations instead of just large audiences, that feels like the real value rather than the AI itself.