Audiary - Text to Speech
效率
仅适用于 iPhone
免费 · App 内购买 · 专为 iPhone 设计。未针对 macOS 验证。
Audiary turns your text into something you can listen to. Paste an article, share a document, or type your own content — then press play and listen with high quality voices powered entirely by on-device AI.
No subscriptions. No ads. No cloud processing. Just clear, natural speech that runs locally on your iPhone.
ON DEVICE NEURAL VOICES
Choose from 20 AI-powered voices with remarkably natural pacing and intonation. Male and female options, each with a unique character. Preview any voice before selecting it. Neural voices run entirely on your device — your text never leaves your phone.
THREE VOICE TIERS
- Neural: Natural AI voices (20 options, one-time 90 MB download)
- Enhanced: Apple's high-quality system voices
- Default: Basic system voices, always available
Pick the tier that fits your device and preference. Switch anytime in Settings.
BUILT FOR LISTENING
- Markdown-aware chapters — headers become skippable tracks
- Adjustable speed with pitch preservation (0.75x to 1.5x)
- Full lock screen and background playback controls
- Word-by-word highlighting in the full-screen player
- Resume exactly where you left off
GET CONTENT IN EASILY
- Paste from clipboard
- Share from any app (text, markdown, PDFs, Notes exports)
- Type or edit directly in the built-in editor
CUSTOM PRONUNCIATIONS
Teach Audiary how to say names, acronyms, or technical terms. Add phonetic spellings in Settings and neural voices will use them every time.
SIMPLE, HONEST PRICING
Audiary comes with 15 minutes of free listening — no account required, no feature restrictions. After that, a single one-time purchase unlocks unlimited playback forever. No subscriptions, no catches.
更多 Plurals and possessives now sound right. Words like "doctor's," "James's," and "countries" were being sounded out as unfamiliar words; now they're spoken from the base word the way they should be. Thanks to the reviewer who pinpointed it!
2.0.1 9月14日
Audiary 2.0 gives the neural voices a second opinion on tricky words:
- Apple Intelligence Assist. Words that are spelled the same but said differently — "read," "bass," "dove," "bow," "wound" — are checked against Apple's on-device Apple Intelligence model whenever the usual rules can't tell which is meant. It runs entirely on your iPhone, works offline, and sends nothing anywhere. On by default on devices with Apple Intelligence; turn it off any time in Settings › Voice.
- The rules got sharper for everyone, on every device: past-tense "read," the bow of a ship versus a bow and arrow, and more. On a public benchmark of these words, mispronunciations fell by about 40 percent.
- Mr., Mrs., Dr., and other titles are spoken as words, so names no longer get a pause in the middle.
Everything still runs entirely on your device. No subscriptions, no ads, no cloud.
2.0 9月11日
Academic papers and reports now sound the way they read:
- Footnotes, superscript reference numbers, and citations like [12] or (Smith et al., 2020) are left out of the audio, so sentences flow instead of stopping for every reference.
- Reference lists, acknowledgements, funding statements, and tables of contents are skipped.
- Equations and table bodies, which never made sense read aloud, are dropped. Table captions stay.
- Page numbers and running headers no longer land in the middle of sentences, and words hyphenated across a line break are rejoined. This helps every PDF, books included.
- Section headings in papers become chapters you can skip between.
- New in Settings: "Skip Citations" also removes reference markers from text you paste. Turn it off if you'd rather hear them.
1.6 9月5日
This update brings your ebooks into Audiary:
- Import EPUB books. Pick one from Files, share it from any app, or tap an EPUB anywhere in iOS and choose "Open in Audiary." Chapters come straight from the book, so you can skip between them like tracks.
- Books are read entirely on your device. Nothing about them is ever uploaded.
- Chapter numbers written as roman numerals. "Chapter IV" is now spoken as numbers by the neural voices instead of letter by letter.
- PDFs can also be opened in Audiary directly from Files and Mail.
Copy-protected (DRM) books can't be read, and Audiary says so plainly rather than failing silently.
Everything still runs entirely on your device. No subscriptions, no ads, no cloud.
1.5 8月28日
This update makes Audiary a proper home for your PDFs:
- Import PDFs right in the app. Pick a file from Files and documents with real text arrive in seconds, with chapters built from the PDF's own outline.
- Scanned PDFs work now too. Audiary recognizes the text page by page, entirely on your device, with progress you can pause and resume. No page is ever uploaded anywhere.
- Book-length imports now show progress during import and editing large documents in app is smoother.
Everything still runs entirely on your device. No subscriptions, no ads, no cloud.
1.4 8月24日
Audiary 1.3 turns shared links into clean, listenable articles.
• Share any web page — Audiary extracts just the article: no menus, ads, or cookie banners, with chapters built from the article's headings
• Edit before you save: the share sheet is now fully editable, with live word count and listen time
• New editor tools: mark chapters with ##, hear how the voice pronounces a selection before saving, find in text, undo and redo
• Chapter controls now work from the first chapter you add
• Fixes and refinements throughout
1.3 8月17日
- Folders. Organize your library — group audiaries into folders, expand them in place, and move items with a long press.
- VoiceOver support. Audiary now works end to end with VoiceOver: every player control is labeled, the scrubber adjusts in 15-second steps, and status updates are announced.
- More accurate word highlighting. The follow-along highlight now stays with the spoken word from the first sentence, not just later in playback.
- Pasted lists keep their shape. Bulleted and numbered lists play as separate, numbered items instead of running together into one paragraph — so you can hear where you are in a list. This also fixes documents already in your library.
- Improved pronunciation. The neural voice now takes sentence context into account for many words that are spelled the same but spoken differently, so more of them land right. You can turn this on or off in Settings.
- Steadier neural voice. Reworked how the neural voice runs so playback stays reliable on more iPhones and iPads, including with the screen locked. On the rare device that can't run it, the app falls back to the built-in voices cleanly instead of interrupting you.
- Still fully on-device. No account, no subscription, no ads.
1.2.5 8月11日
- Improved pronunciation. The neural voice now takes sentence context into account for many words that are spelled the same but spoken differently, so more of them land right. You can turn this off in Settings.
- Steadier neural voice. Reworked how the neural voice runs so playback stays reliable on more iPhones and iPads, including with the screen locked. On the rare device that can't run it, the app now falls back to the built-in voices cleanly instead of interrupting you.
- Still fully on-device. No account, no subscription, no ads.
1.2.4 7月28日
Under-the-hood updates to keep neural voices running smoothly.
1.2.1 7月10日
Neural voices now install faster and run much lighter — a smaller download and a big drop in memory use, so on-device AI speech runs comfortably on more iPhones.
Also in this update:
- Smoother playback on long articles and documents — no more pauses or clipping in long passages
- More accurate reading of numbers, years, dates, and hyphenated words
Still 100% on-device. One-time purchase — no subscription, no ads.
1.2 7月9日
• Voice model now downloads in the background — you can leave
the app or lock your phone and it'll keep going.
• Smoother playback over Bluetooth and CarPlay, plus a fix
for clipping at the start of a session.
• Polished the onboarding flow.
1.1 5月21日
Plurals and possessives now sound right. Words like "doctor's," "James's," and "countries" were being sounded out as unfamiliar words; now they're spoken from the base word the way they should be. Thanks to the reviewer who pinpointed it!
更多 版本 2.0.1 9月14日
未收集数据 开发者不会从此 App 中收集任何数据。