I’m experimenting with a browser that lets people comment directly on web pages
It’s a small experiment around what happens when comments live with content instead of platforms.
I started building it because I often felt that conversations about content were scattered across platforms and lost their context.
It’s still early, and I’m honestly unsure where the real value is yet.
I’d love feedback from people who build products or care about UX:
• Does this feel useful or noisy?
• When would you not want comments on a page?
• What feels confusing or uncomfortable about this idea?
If it helps, here’s a public beta you can try:
https://testflight.apple.com/join/1JAU8yNP (TestFlight)
Even a quick gut reaction would be helpful.
Interesting approach — the core tension you're playing with is whether context belongs with content or with communities. Right now, comments live on platforms (Twitter threads about articles, HN discussions, etc.), which fragments conversation but also creates moderation/quality signals.
A few questions that might help you find where the value is:
When it's useful:
When it might get noisy:
The hard question:
What makes someone come back to check comments on a random webpage? That's the distribution problem — even if the feature is useful, you need a reason for people to remember to use it.
Curious what your initial usage patterns look like — are people mostly commenting on articles, product pages, or something else?
This is a really fair framing — especially the distribution problem.
Right now I’m less confident about where this works at scale, and more interested in observing where it breaks first.
Early signals (very limited so far) skew toward articles rather than product pages, but it’s too early to call.
Rather than expecting people to remember a specific page, I’m experimenting with whether the pull can come from elsewhere:
– seeing where conversations are happening across the web (a trends view)
– having your own past comments act as a memory trail
So the habit isn’t “check comments on this page again,”
but “check where conversations are happening, and where I’ve participated.”
I agree that public pages without quality signals are the hardest case — that tension is probably the core of the experiment.
Whether this is enough to create a return loop is still very open, and honestly one of the things I’m most curious to learn from early usage.
The "trends view + memory trail" reframe is smart — you're essentially building a second feed layer on top of the open web.
That shifts the value prop from "comments on pages" to "discover conversations happening around content you care about." Much stronger pull.
The memory trail piece is interesting too — it turns commenting from a one-shot action into building your own context graph over time. That could compound.
One thing I'd watch: how do you handle the "cold trends" problem? If trends only show up after critical mass, early users won't see much activity. Seeding with curated sources (top tech blogs, etc.) might help bootstrap the network effect.
Good luck with the experiment — curious to see what patterns emerge.
I like the idea, it's fresh and interesting. Good luck!