Voice-to-Text in Any Language, Free & Offline
Published 26 June 2026 · ~6 min read
VoxFlow is free — no account, no subscription, on-device.
Most dictation tools assume you speak English
If your first language isn't English — or you switch between two of them mid-sentence the way most bilingual people do — voice-to-text tends to let you down. Built-in dictation supports your language in some regions but not others, mangles names and technical terms, and the moment you lose signal abroad, cloud transcription apps simply stop working. You end up typing anyway, which is exactly the friction you wanted to avoid.
There's a better setup that most people don't realise is now possible on an ordinary iPhone: speak in almost any language, fully offline, and get a clean summary in Apple Notes — for free, with no account. The reason it works comes down to where the speech model runs.
Why on-device Whisper handles ~99 languages
VoxFlow runs WhisperKit, an Apple-silicon-optimised port of OpenAI's Whisper model, directly on your device. Whisper was trained on a huge multilingual dataset and recognises roughly 99 languages — not just the big European ones, but Arabic, Hindi, Mandarin, Vietnamese, Tagalog, Swahili and many more. It's also unusually good at accents and at code-switching, where you drop a few English words into a Slovak or Spanish sentence the way people actually talk.
Because that model lives on your phone, two useful things follow at once. It costs the developer nothing per transcription, so there's no subscription to recover server bills — and it needs no connection, so it keeps working when you're offline. The same property powers our guide on why on-device voice-to-text can be free.
The workflow: speak your language → summary in Notes
The loop is the same whatever language you think in:
- Long-press the Apple Watch crown (or hit record on iPhone or Mac) — no menu-digging while you're walking or travelling.
- Talk for 30 seconds to a few minutes in your own language — an idea, a to-do list, a recap of a conversation. Mix in English terms freely; the model copes.
- It transcribes on-device, offline, wherever you are — a flight, a metro abroad, a village with no coverage.
- A clean summary plus key points lands in Apple Notes — in the language you spoke, ready to share or act on.
It's the same capture mechanic behind turning Apple Watch voice memos into action items and capturing meeting action items in 30 seconds — just unbounded by language.
Multilingual voice-to-text options compared
How the realistic choices stack up in 2026 for dictating notes in your own language:
| Tool | Price | Languages | Works offline? | Summary out? |
|---|---|---|---|---|
| Apple Dictation | Free | ~30, region-locked | Some, on newer devices | No — raw text |
| Otter.ai | ~€170/yr | English-first | No (cloud) | Yes |
| Google / cloud STT apps | Varies | Many | No (cloud) | Sometimes |
| Generic AI note apps | ~€100–200/yr | Many (cloud) | No | Yes (account) |
| VoxFlow | Free | ~99 (Whisper) | Yes, fully | Yes |
The cloud tools cover plenty of languages, but they need a connection, a subscription, and an account — and they upload your voice in whatever language you speak. For the everyday need — "just let me think out loud in my language and drop the result in Notes" — an on-device, offline, free option does it without any of that. If you're leaving a paid recorder, see the best Otter alternative with no subscription.
Where this matters most
Three situations where multilingual, offline dictation quietly wins: travelling — capture ideas and to-dos abroad with no roaming data and no signal in a tunnel or on a plane; bilingual work — switch between your native language and English in a single note without the tool falling apart; and privacy-sensitive notes — client conversations, health, or personal thoughts stay on your device, never uploaded to a server in any language, with no account attached. If you study in a second language, the same loop turns lectures into study notes you can actually revise from — and if you write in more than one language, it can turn a spoken idea into a first draft for your newsletter or show notes.
Frequently asked questions
Is there a free voice-to-text app that works in any language?
Yes. VoxFlow runs Whisper on-device via WhisperKit (~99 languages) and is free with no subscription and no account. You speak in your language and get a transcript plus summary in Apple Notes on iPhone, Mac and Apple Watch.
Can voice-to-text work offline without internet?
With an on-device app the transcription runs locally, so it works with no internet at all — a plane, abroad without data, or Wi-Fi down. Cloud tools stop the moment you lose signal.
Does Apple dictation support as many languages as Whisper?
Apple dictation covers a few dozen languages and gives raw text only, with region-dependent support. Whisper recognises around 99 languages and handles accents and code-switching, then VoxFlow turns it into a summary in Notes.
Is it private to dictate notes in another language?
Yes — on-device means your audio never leaves your device in any language, with no account and no cloud copy. That matters most for sensitive notes and client conversations.
Try it
VoxFlow is live on the App Store now — iPhone, Mac, and Apple Watch native, on-device Whisper across ~99 languages, completely free with no subscription, no in-app purchases, and no account. Think out loud in your own language, anywhere, and leave with the summary already in Notes.
→ Download VoxFlow on the App Store
More guides: all VoxFlow voice-notes guides.
VoxFlow is built by an independent developer in Bratislava. On-device, free, no account — because the language you think in shouldn't decide whether your notes tool works.