Pagelive Review: Private Links for AI-Built Pages, Priced Like a Side Bet
Pagelive does one narrow thing: it takes an HTML page, usually one an AI just spat out, and puts it behind a link that Google will never index unless you explicitly ask. That is the whole pitch, and to its credit, the product does not pretend to be anything else.
The problem it solves is real. Claude and friends can now build a genuinely good-looking proposal or dashboard in one shot, and the obvious ways to share it (GitHub Pages, Netlify Drop, Vercel) all default to public and indexable. Fine for a personal site. Less fine for a client pricing proposal. Pagelive's homepage makes this concrete with a demo of a test page indexed on a "typical free host" within a day, and the claim checks out mechanically: pages ship with robots meta and X-Robots-Tag headers, and the FAQ confirms indexing is opt-in per page.
What you actually get
Publishing works three ways: drag-and-drop an .html file (no sign-up needed for the first one), a REST API, or an MCP connector for Claude. The MCP route is the slick one. Connect once at app.pagelive.io/api/mcp, say "publish this as a Pagelive page," and Claude ships the HTML and hands back the link. One catch the marketing underplays: custom MCP connectors in Claude require a paid Claude plan, so the headline workflow is not available to free-tier Claude users.
Once a page is live you get viewer analytics: views, unique viewers, referrer, dwell time, geography, and a per-day history, with bot opens filtered out and an email alert the moment someone opens it. Pages can carry password protection, email gates, and reply forms for lead capture, and updating a page never changes the URL. Free links live on a random pagelive.site subdomain (think quiet-harbor-42.pagelive.site), which looks fine for internal use but a bit throwaway on a client proposal.
The pricing is the interesting part
Three plans. Free is genuinely usable: 5 active pages, basic analytics, one password-protected page, no credit card. Founding is the full product for $7/mo (list price $15), "locked for life," with unlimited fair-use pages, 5 custom domains, white-label badge removal, and the full analytics depth. Team is $19/seat/mo with a 2-seat minimum, 25 domains, and shared workspaces.
Two things bug me here. First, the "85 of 100 founding seats left" scarcity framing. This is a young product with no visible traction, and seat-count urgency on a lifetime-locked price is a pressure tactic, not a discount. Second, the pricing page has a Monthly/Annual toggle but no annual prices published on the captured page, so if you want to commit longer you cannot even see the number.
What nobody is saying yet
And that is the honest con: almost nobody is saying anything. Searching for independent sentiment turns up directory listings (Slashdot, SourceForge comparisons, a couple of MCP server catalogs like LobeHub) and Pagelive's own comparison pages against Papermark and own.page, but no organic Reddit threads, no Product Hunt comment sections full of real users, no "I switched from X to Pagelive" posts. The SOC 2 Type II and EU data region claims are infrastructure-level (it is built on Cloudflare), not an independent audit of Pagelive itself, and I could not find third-party verification of either. For a tool you might point confidential client work at, that absence of track record matters more than any feature gap.
Pick it or skip it
Pick it if you regularly build one-off pages in Claude and need to send them to clients without leaking them to search engines. The free tier covers a real evaluation, and $7/mo for unlimited pages with custom domains and open-tracking is cheap for what agencies pay for DocSend-style analytics elsewhere.
Skip it if you need social proof before trusting a vendor with client-facing links, if your whole team is on free Claude plans (the MCP connector is gated), or if you are a solo operator tempted by Team: at $19/seat with a 2-seat floor, you are paying $38/mo minimum for workspace features a one-person studio will never touch. For everyone else, this is a $7 experiment with a free tier underneath it. Low risk, real utility, unproven company.