Just launched OriginBrief on Product Hunt — continuous AI research from primary sources
Hey IH 👋
I just launched OriginBrief on Product Hunt yesterday. It's a continuous research SaaS that monitors primary sources (company blogs, government sites, research orgs) for the topics you care about, and delivers weekly AI-generated reports.
The "why":
After spending too many hours every week chasing news aggregators that just repeat each other, I wanted something that focuses on signal over volume — primary sources only, structured weekly reports.
Stack: Next.js 15 + Supabase + Claude API + Vercel cron
Pricing: Starter $33/mo · Pro $65/mo (PH coupon "PRODUCTHUNT" gives 50% off)
Would love feedback from this community — what topics would you track?
Congrats on the launch, Aritomo! You’ve hit a real pain point—most news aggregators are just echo chambers.
Since you’re already focusing on primary sources, have you thought about positioning this for Journalists or Digital PR agencies? They are constantly looking for 'Signal over Noise' to find original stories. If OriginBrief can help them spot a trend before it hits the mainstream media, you've got a high-ticket enterprise use case right there.
Actually, I work in the PR and media placement space, and I can tell you that getting featured on top-tier news sites often starts with exactly the kind of primary data your tool provides. Would love to chat about how you're planning to scale the distribution for this!
Thanks Muhammad — the journalist/PR angle is one I hadn't explored, and your "spot a trend before it hits mainstream media" framing is sharper than how I'd been thinking about it. The primary-source-first approach does map to that use case naturally. Really appreciate the industry read — will sit with this.
Glad it resonated! Primary-source monitoring is genuinely underused in PR — most teams are still stuck on Google Alerts and hope for the best.
If you ever want to explore the journalist/agency angle properly, happy to have a quick chat. No pitch — just curious whether there's something worth building on together.
Thanks Muhammad — I'd really love to explore this further.
One honest caveat: spoken English is still a challenge for me,
so I'd have to pass on a live call. But I'm more than happy to
go deep over email, at whatever pace works for you. Async also
gives me time to think properly, which I appreciate.
Heads-up on timezone too — I'm based in Japan (JST), so replies
might come with a 12–15 hour lag depending on your side. Not
ignoring you, just sleeping.
If that works for you, feel free to reach me at
aritomo.fukuda@couch-potato.co.jp. Genuinely curious to hear
more about how your side of the industry actually works.
Thanks again for the thoughtful framing — it's been sitting
with me since you posted it.
Thanks, Aritomo! Just sent you an email with some thoughts on the PR/Journalist use case. Let's catch up there!
Love the ‘signal over volume’ angle — especially focusing on primary sources, that’s a real pain for anyone doing serious research.
Curious though — once someone identifies valuable signals, how are you seeing them validate which insights actually translate into real decisions or demand? That jump from insight → outcome is where most tools lose people.
I’ve been seeing some founders run small, high-intent experiments (fixed low entry, capped spots, strong upside) alongside tools like this to quickly test what actually matters to users — surprisingly effective for early traction.
Feels like that layer could complement OriginBrief really well. Have you explored something like that?
Great question — validation of "did this signal actually matter" is the part we haven't solved yet. Right now we optimize for surfacing what's new/changed, not for scoring impact. The small-experiment pattern you described is exactly the kind of layer I think would complement this — noted, and will think about it more. Thanks for the sharp framing.
Love that — surfacing change first is a solid foundation.
Impact scoring usually gets clearer once you tie it to a specific decision or action taken after the signal (even something simple like “did this trigger a change?”).
The small-experiment layer can sit nicely on top of what you already have — not replacing it, just giving it feedback loops.
If you end up exploring that, would be interesting to compare notes 👍
honestly this is something I'd actually use. got so tired of reading the same five versions of every story each week, all copying each other.
arxiv would be huge for me. there's like 200+ papers a week in my field and I already burn 2 hours just figuring out which ones matter. if something could synthesize that I'd pay for it no question.
not sure if you've thought about this but some kind of "diff" view would be cool too, like what changed in a source vs last week. sometimes "nothing new here" is actually the useful information.
anyway congrats on launching, good to see this kind of stuff actually getting built
"nothing new here is actually the useful information" — this might be the best reframe I've read on this. Stable = silence is wrong; stable = verified-unchanged is the real value.
And yes, arxiv is on the list. 200+ papers/week is exactly the shape of problem this format fits — will build toward that.
This looks really useful — especially the focus on primary sources instead of recycled content.
I like that you don’t end up getting the same report twice, that’s usually the biggest issue with this kind of tooling.
Thanks Dan. The "same report twice" problem is specifically what the change-tracking layer is trying to fix — so if you end up trying it, I'd really want to hear whether that part holds up in practice.