Voice to Text: Dictation and Voice Messages to Text (2026 Guide)

Voice to text is simply speech converted into written words, either as you talk (dictation) or from a recording you already have (a voice message, a voice memo, a call). The fastest route depends on which of those you need: your phone and computer already include free dictation, while voice messages and longer notes are better handled by a transcription tool that takes a file. This guide walks through all three, with the honest limits of each.

Three kinds of "voice to text", three different tools

People search for "voice to text" meaning quite different things:

1. Live dictation. You speak, text appears in a document, message or email. Built into iOS, Android, Windows and macOS.
2. Voice messages. Someone sent you a WhatsApp or Telegram audio and you would rather read it. Some messengers can transcribe in-app; otherwise you export the file.
3. Voice notes and recordings. Memos, lectures, calls or interviews recorded earlier. These need a transcription service that accepts files and can handle length, multiple speakers and timestamps.

The distinction matters because dictation engines are tuned for one person speaking clearly into a nearby microphone, in short bursts, with you watching the screen. A twenty-minute voice memo recorded in a car is a different problem, and the tools that solve it well are different too.

Dictation on your phone: iPhone and Android

iPhone and iPad. Tap the microphone icon on the keyboard and start speaking. Punctuation can be spoken ("comma", "full stop", "new paragraph") and recent iOS versions add some punctuation automatically. Dictation works in any text field, including Notes, Mail and messaging apps. For supported languages it runs on the device, which is good for privacy, and it keeps working without a connection.

Android. Gboard's microphone button does the same job. Google's voice typing is generally strong for English and the major European languages, and on recent Pixel phones the Assistant-powered voice typing adds automatic punctuation and voice commands. Samsung's keyboard has an equivalent button. Google's Recorder app on Pixel devices transcribes recordings live, which is a useful middle ground between dictation and transcription.

Phone dictation is excellent for messages, quick notes and short emails. Its weak spots are long passages (the engine sometimes stops or drops a sentence while you pause), technical vocabulary, and anything with more than one person talking.

Dictation on desktop: Windows, macOS and the browser

Windows 11. Press Win + H in any text field to open Voice Typing. It supports automatic punctuation (in its settings) and a handful of commands such as "delete that". It needs an internet connection for most languages.

macOS. Enable Dictation in System Settings → Keyboard, then press the shortcut (by default the Microphone key, or pressing Control twice, depending on your Mac). On Apple-silicon Macs dictation runs largely on device and can run for long stretches.

Google Docs. In Chrome, open Tools → Voice typing. This is convenient if you write in Docs anyway and it supports many languages, but it only works while the Docs tab is active and in front.

Browser-based transcription. If you want dictation that produces a proper editable transcript rather than text in a field, record in a web app instead. Audiotexta's web app, for example, records from the microphone, transcribes, and lets you edit with the audio playing back, which makes it easier to fix a word you cannot remember saying. The browser extension gives you the same from any tab.

Converting WhatsApp and Telegram voice messages to text

Voice messages are the most common reason people look for voice to text, so let us be specific.

Built-in options

• WhatsApp has rolled out voice message transcripts that are generated on your device. You enable them in Settings → Chats → Voice message transcripts, choose a language, then long-press a message and tap Transcribe. Language coverage is limited and the feature is not available on every platform or version, so check whether the option appears for you.
• Telegram can transcribe voice and video messages with a tap on the "→A" icon next to the message. It is a Telegram Premium feature with a small free allowance for other users.
• iMessage transcribes audio messages automatically on recent iOS versions.

If those work for you, use them. They are the fastest route, and the audio either stays on the device (WhatsApp) or is handled by the messenger's own servers (Telegram).

When you need more

Built-in transcripts tend to fall short when the message is long, in a mixed language, recorded in a noisy place, or when you need to save or search the text later. In that case export the file:

• In WhatsApp, long-press the voice message → Share (or Forward) → save to Files or a cloud drive, or send it to yourself by email. Files are usually Opus audio in an .ogg or .opus container, sometimes .m4a on iPhone.
• In Telegram, open the message menu → Save to Downloads / Save to Files. Telegram voice messages are .ogg (Opus).

Then upload the file to a transcription service. Most modern services accept Opus and OGG directly; if yours does not, VLC or a desktop converter can turn the file into MP3 in seconds. In Audiotexta you can drop the file into the web app, get the text with timestamps, and export to TXT or DOCX. The free plan covers a reasonable number of short messages per month; see pricing for the current limits.

A practical tip: when you receive a batch of voice messages from the same person, forward them all to yourself and transcribe them together. You get one searchable document instead of a dozen fragments.

Transcribing voice notes and memos you recorded earlier

Voice memos from Apple's Voice Memos app arrive as .m4a; Android recorders typically produce .m4a, .mp3, .3gp or .amr; dedicated recorders often use .wav. Any of these can be uploaded to a transcription tool, and the workflow is the same as for voice messages:

1. Get the file onto the device where you will transcribe (AirDrop, cloud drive, email to yourself).
2. Upload it, choose the language, and enable speaker labels if the note includes a conversation.
3. Use the AI summary or key-points option if the note is long and you only need the gist.
4. Export in the format you need: TXT for pasting, DOCX for editing, PDF for sharing.

For long notes, timestamps earn their keep: you can jump back to the moment you said something ambiguous and check what you meant. For a deeper walk-through of file formats and limits, see our guide on transcribing audio to text; recordings that happen to be video are covered in transcribe video to text.

Tips for clean dictation and better voice-to-text results

Whatever tool you use, a few habits make a large difference:

• Get the microphone close. A phone held near your mouth beats a laptop microphone across a desk. Earbuds with a microphone are a good compromise.
• Speak in complete thoughts. Dictation engines use the surrounding words to pick the right one; half-sentences give them less to work with.
• Say punctuation when dictating; do not bother when transcribing. Live dictation needs "comma" and "new line"; file-based transcription adds punctuation for you.
• Spell unusual names once, then correct them in the editor. Most engines will not know your colleague's surname or your product name.
• Avoid background audio. Music, television and a running tap are the most common accuracy killers.
• Choose the language explicitly if you know it. Auto-detection is good but not perfect on short clips.
• Proofread. Even with clear audio, expect the occasional wrong homophone. Reading the text while the audio plays is the fastest way to catch them.

There is more on what actually drives accuracy in our accuracy tips.

Which approach should you use?

Privacy: where does your voice go?

On-device dictation (iOS, Apple-silicon Macs, some Android models) keeps audio on the phone or computer. Cloud dictation and transcription send audio to a server, so it is worth knowing where that server is and how long audio is kept. If you handle personal data or work under GDPR, prefer services that let you delete recordings, state their retention policy clearly and do not train models on your audio. Audiotexta lets you delete files after transcription and does not use your content for training; our GDPR guide explains what to check with any provider.

Voice to text is no longer a novelty feature. For short text, use what is built into your device. For everything you want to keep, search or share, treat the recording like any other audio file and transcribe it properly.