How to share Perplexity output as a web page

Why not just send the Perplexity link?

Because a Perplexity share link shares the thread — the query, the follow-ups, the whole conversation — not a clean standalone page. And a raw .md or .txt file isn't a page either: the person you send it to has to download it and open it in something. Neither is a URL you can drop in a message and know it renders the same for everyone.

A hosted page is. Copy the one answer into a single file, publish it, and it becomes a web page served like any other — with the reading captured on the other side.

How to publish Perplexity output

The loop is short:

1. In Perplexity, take the answer or research report you want and copy it — with its citations — as Markdown into one self-contained file. One file, not the thread link.
2. Paste it, or drop the file, into the composer at ilolink.com.
3. Pick a visibility mode — public, unlisted, private (teamspace members only), password, or expiring.
4. Get ilolink.com/<slug> and send it.

The branded link 302-redirects to an isolated render origin, view.ilolink.com, where the doc is served under a strict content-security policy (default-src 'none'). Markdown renders straight to clean HTML, so a Perplexity answer is a natural fit — headings, lists, and links come across as written. If you paste HTML instead, by default it's sanitized on the way in: scripts are dropped, so any interactive JS is frozen to its static state while the layout and CSS render. If you mark the doc as trusted at publish time, it's kept as-is and its own scripts run inside a sandboxed frame on the isolated origin. Keep styling as inline CSS and images as absolute https: or data: URLs — relative asset paths and http: images won't load. The cap is 15 MB per doc. If you started from raw Markdown, see turning Markdown into a web page.

What you learn after sharing

This is the part the thread link never gave you. Once the page is out, ilolink shows how it was actually read — all cookieless, from a rotating visitor hash, with no fingerprint and no personal profile:

• Views — total, plus approximate unique views (uniques are approximate by design).
• Scroll funnel bucketed at 0 / 25 / 50 / 75 / 100%, so you see where people stop reading the report.
• Heatmaps — click and scroll, viewable by device class.
• Reactions and notes — readers can react or leave a short note, no account.
• Comments — threaded, anchored to the doc.

There's more — average time on page, top referrers, countries, device class, 30-day daily views. See analytics, heatmaps, and feedback for the full picture, or open your documents once you've published one. This page is one of several source guides under sharing AI output.