Features

Everything between a link and a listen

StarCast is a full pipeline: it fetches your content, writes it into a podcast, voices it — in your own voice if you want — and gives you a proper place to listen, on the web or in the podcast app you already use.

Start from anything

Four kinds of input, one paste box. StarCast extracts the substance and leaves the clutter behind.

Article URLs

Paste a link to a blog post, news story, or essay. We pull the readable text — no ads, no nav, no cookie banners.

YouTube & Bilibili

Drop a video link and we work from its subtitles or transcript, so an hour-long talk becomes a tight audio episode.

PDF documents

Upload papers, reports, or e-book chapters. Text PDFs are parsed directly; scanned pages go through OCR.

Pasted text

No link at all? Paste raw text — meeting notes, a newsletter, a draft — and turn it into an episode.

Three formats, one source

Pick how you want to hear it. You can regenerate the same source in another format without re-fetching anything.

Signature

Two-host dialogue

Two AI hosts unpack the content together — questioning, summarizing, and connecting the dots. The Deep-Dive style that makes dense material genuinely listenable.

Solo explainer

One voice rewrites the source into a natural spoken explainer. High information density, clear structure, no filler.

Verbatim read

A faithful reading of the cleaned original text. Nothing added, nothing rephrased — ideal when the author’s exact words matter.

A listening experience built for it

Generated episodes land in your personal library, with a player designed for spoken content — not a generic audio tag.

Personal library

Every episode you generate is kept, searchable, and replayable — your private feed of everything you meant to read.

Play queue with resume

Line up episodes and listen through. Stop mid-episode and StarCast picks up exactly where you left off, on any device.

Speed control

Listen from 0.5× to 3×. Your preferred speed is remembered across episodes.

±15s skip

Jump back fifteen seconds to re-hear a point, or ahead to skip what you already know.

Transcript view

Read along while you listen, or skim the transcript afterwards to grab a quote or reference.

Lock-screen controls

Play, pause, and skip from your phone’s lock screen — the player behaves like a real podcast app.

Your private RSS feed

Prefer your own podcast app? Every account gets a private RSS URL — subscribe once, and new episodes appear there automatically.

Works in any podcast app

Apple Podcasts, Overcast, Pocket Casts, AntennaPod — anything that accepts an RSS URL can carry your feed. Xiaoyuzhou does not support custom RSS, so use one of the others for your private feed.

Tokenized and private

The feed URL contains a secret token. Only people with the link can access it, and it never appears in public directories.

Reset any time

Shared the link by mistake? Reset the token in settings and the old URL stops working immediately.

Voices you choose — including your own

Host your feed with a curated preset, or with a clone of your own voice. Cloning is live today: one short sample and every episode is read by you.

Voice picker

Choose a distinct voice per host in dialogues, or a single narrator for explainers and reads. Eleven presets are included on every plan.

Clone your own voice

Upload or record a 10-second to 5-minute sample, confirm you are entitled to use that voice, and your clone hosts every format from then on. Available on Pro and above.

Nine output languages

English, Chinese, Spanish, Japanese, French, German, Korean, Portuguese, Russian. Episodes follow the source language by default, or cross over to the one you pick.

Honest billing

You pay for audio you actually get — nothing else.

1 credit = 1 minute

Pricing you can compute in your head: a 12-minute episode costs 12 credits. No hidden multipliers.

Charged only on success

If extraction or generation fails, you pay nothing. Credits are deducted only when a playable episode lands in your library.

30 free minutes to start

Every new account starts with 30 minutes of generation — enough to try all three formats before spending anything.

Hear it for yourself

Paste your first link — the first 30 minutes are free.

Open the studio