Beyond text-to-speech

Text-to-speech for articles, without the robot

Instead of narrating a web page line by line, Bookmark.fm rewrites it as a conversation between a host and an expert, then voices both sides.

Why straight narration is hard to listen to

Writing is built for eyes. It leans on headings, layout and the ability to glance back a paragraph. Read aloud verbatim, all of that disappears — which is why a narrated article takes more concentration than the original, not less. A conversation carries its own structure: someone asks, someone explains, someone pushes back. That's a format your ears already know how to follow.

How it works

  1. 1

    Save a link

    Paste a URL, click the bookmarklet in your browser toolbar, or post it to the API.

  2. 2

    We write and voice the episode

    Bookmark.fm reads the page, scripts a host-and-expert conversation about it, and records both voices.

  3. 3

    Listen wherever you listen

    The episode lands in your private podcast feed, ready in your usual app within minutes.

What you get

Two voices, clear roles

A host with the questions, an expert with the answers. You always know who's talking and why.

Shorter than the source

Boilerplate, navigation cruft and repetition don't survive the rewrite. You get the substance.

No page prep

Bookmark.fm extracts the article text itself, including on many sites that block simple scrapers.

Delivered as a podcast

Episodes arrive through a private RSS feed, so playback, download and speed controls come from the podcast app you already use.

Transcript included

Prefer to read a section again? The written transcript sits on the episode page next to the original link.

Frequently asked questions

Is this verbatim text-to-speech?

No, and that's the main difference. If you need the exact words of the article read aloud — for study, proofreading or accessibility reasons that require fidelity — a dedicated TTS reader is the right tool.

Can I change the playback speed?

Yes, in your podcast app. Episodes are ordinary MP3 files, so every player's speed and skip controls work.

Which voices does it use?

Two distinct AI voices, one per speaker, chosen so the host and expert are easy to tell apart.

Does it work on any URL?

Most public web pages with article text work. Pages needing a login, and content that's purely video or PDF, don't.

Start listening to your reading list

Save your first link and have an episode waiting in minutes.