Sayble: AI Voice Keyboard

Sayble: AI Voice Keyboard

Speak freely. Write clearly.

0 ratings
Free
In-App Purchases

Rating summary

Details

Released
August 22, 2026
Updated
October 29, 2026

Features

Sayble: AI Voice Keyboard screenshot #1 for iPhone
Sayble: AI Voice Keyboard screenshot #2 for iPhone
Sayble: AI Voice Keyboard screenshot #3 for iPhone
Sayble: AI Voice Keyboard screenshot #4 for iPhone
Sayble: AI Voice Keyboard screenshot #5 for iPhone
Sayble: AI Voice Keyboard screenshot #6 for iPhone
Sayble: AI Voice Keyboard screenshot #7 for iPhone
Sayble: AI Voice Keyboard screenshot #8 for iPhone
Sayble: AI Voice Keyboard screenshot #9 for iPhone
iphone
iphoneDuo
ipad
🖼️Get Icon
Icons↘︎

About

Speak the way you think. Sayble turns spoken words into text you can send without editing. Dictation gives you words. Sayble gives you finished sentences: the "um" and "uh" are gone, the stutters are collapsed, the sentence you restarted halfway is resolved, and the punctuation is there. It works inside your keyboard, so it works everywhere you type. WHAT IT FIXES • Filler words — "um", "uh", "you know", 嗯, 呃 • Stutters and repeated words — "I I I think" • Mid-sentence corrections — say "three o'clock, no, four" and you get four • Missing punctuation, including commas and question marks • Run-on speech, broken into readable sentences and paragraphs • Spoken lists — "first... second... third" becomes a numbered list YOU CHOOSE HOW MUCH IT CHANGES • Verbatim — your exact words, punctuation only • Smart — removes filler, stutters, and corrections • Polished — also improves phrasing and tone No mode invents facts, numbers, or commitments you did not say. REWRITE WHAT'S ALREADY THERE Select text and pick an action: Proofread, Concise, Formal, Friendly, or Bullets. Or say what you want changed. Every rewrite shows a preview first, and Undo restores the original. WRITE IT FROM SCRATCH Empty field? Say what you need — "tell my colleague I'm off tomorrow" — and get the finished message, ready to send. CLEAN UP WHILE YOU TYPE Optional: fillers, stutters, and duplicated words disappear shortly after you pause. Wording, punctuation, and numbers are never changed silently, and the word you're still typing is left alone. SNIPPETS Save addresses, signatures, and standard replies once, insert them from the keyboard with one tap. PERSONAL DICTIONARY Add names, product names, and technical terms so they come out right. You can see and delete every entry — nothing is learned silently. CHINESE AND ENGLISH Built for both, including mixed input. Spacing between Chinese and Latin text is handled automatically. Interface language and recognition language are independent. PRIVACY • No audio recorded. Keyboard dictation uses Apple's built-in iOS dictation, and the Sayble keyboard does not request microphone access • Only text is processed • No account, no sign-in, no email • No ads, no third-party analytics, no tracking • History, snippets, and dictionary stay on your device, and history can be cleared anytime • The original transcript is always recoverable REQUIREMENTS The keyboard needs Allow Full Access for cloud refinement and for sharing snippets with the app. Password fields always use Apple's secure keyboard — iOS does not allow third-party keyboards there. Privacy Policy: https://cloudsjhan.github.io/sayble-privacy-policy/ Support: https://cloudsjhan.github.io/sayble-support/
Show more

What's New in Sayble

2.0.0

October 29, 2026

Sayble 2.0 lets you choose which models turn your voice into text: Sayble's cloud, AI services you already pay for, or models that run entirely on your iPhone. • Bring your own models: add an API key from Alibaba Cloud Bailian, DeepSeek, Gemini, OpenAI, Kimi, Zhipu or MiniMax, or connect any OpenAI-compatible service, including Ollama or LM Studio on your network. Requests go straight from your iPhone to the provider and don't count toward your Sayble allowance. • Three ways to recognize speech: Direct (an audio model listens to the recording), Hybrid (speech-to-text, then a text model cleans it up), and Offline (an on-device model, nothing uploaded). • On-device recognition with SenseVoice Small or Qwen3-ASR 0.6B. Downloads resume after an interruption and are verified before use. • Voice translation: speak in one language and write in another, in 16 languages. Change the target language from the home screen or the keyboard, even while recording. • Translate selected text in the keyboard, then copy the result, try again, or translate it into another language just this once. • Recommended setups: apply a complete configuration in one tap, see its estimated monthly cost, and test it before you switch. • Track your spending for today, this month or all time, broken down by model, with a trend chart. Keys are checked before they're saved. • Your dictation at a glance: words dictated, time saved, speaking rate, streaks and a daily activity grid. • History is grouped by day and can be filtered by dictation, voice translation or selection translation. Translations show their original text, and failed attempts are kept with the reason. • Search your dictionary, or import terms from a CSV or text file. • Fixes and refinements: Sayble now asks before clearing history or spending records, or deleting a key or a model. History has larger tap targets, models appear by name instead of internal IDs, the home screen's engine summary fits at large text sizes, and the keyboard's AI key shows its icon again when cloud AI is off.

More
3

Subscriptions & In-App Purchases

Sayble Pro Monthly

Unlimited AI writing for one month.

$0.991 Month
How to cancel Sayble subscription ⁉️

Developer apps