Speech to text

Turn recorded speech into accurate text

Upload a recording of anyone speaking and get a speaker-labeled, timestamped transcript in minutes. 95%+ accuracy on clear audio, 90+ languages detected automatically, files up to 20 hours. The first 10 minutes of every file are free, no credit card.

Convert speech to textFirst 10 minutes free. No credit card.

Spoken message in, readable text out

A sample voicemail and its transcript, synced. Press play, or click any word to jump to that moment. Your file works the same.

Sample preview
play it · click any word to jump

Voicemail:Hey, it's Jordan calling from Riverside Dental. Just following up about your appointment next Tuesday at two. If you need to reschedule, give us a call back at the office. Otherwise, we will see you then. Thanks. Bye.

0:00
0:13

What is speech to text?

Speech to text is technology that converts spoken words into written text using speech recognition. It comes in two forms: live dictation, which types while you talk, and file transcription, which turns an existing recording into a document. FileToText does the second: upload a recording of speech, in any of 90+ languages, and get back an accurate transcript with speaker labels and timestamps that stays private to your account.

Recordings, not live dictation

Honest scoping so you pick the right tool the first time.

Typing while you speak?

That is live dictation, and you already have it for free: the microphone key on your phone keyboard, or voice typing in Google Docs. Great for composing a message, but it cannot open an existing recording.

Already have the recording?

That is FileToText. Upload the file after the fact and get the full transcript: multiple speakers labeled, timestamps on every word, and exports your dictation app cannot produce.

How it works

1. Record anywhere

Phone voice memo, meeting recording, dictated note, saved call: any device works. FileToText reads MP3, WAV, M4A, AAC, OGG, FLAC, and most video formats.

2. Upload the file

Drag the file in and it uploads directly, no conversion step. The spoken language is detected automatically from 90+ options.

3. Get verified text

A speaker-labeled, timestamped transcript synced to the audio. Click any word to hear that moment, then export to TXT, DOCX, SRT, or WebVTT.

Voice to text, for the recordings phones make

Most voice-to-text jobs are short personal recordings. Test the transcript quality free on the first 10 minutes of every file.

Voice memos to text

iPhone and Android memos, speaker-labeled and timestamped.

Voicemail to text

Read the message instead of replaying it.

M4A to text

The format phones actually record in.

Interview to text

Two voices, labeled, every quote checkable.

Related tools

Thinking in files rather than speech? The audio to text converter covers every format. Curious how the AI itself works? See AI transcription, or browse everything you can transcribe.

Speech to text FAQ

Updated July 2026
To convert recorded speech to text, upload the recording to FileToText: an MP3, WAV, M4A, voice memo, or even a video file. The speech recognition engine transcribes it in minutes, detecting the spoken language automatically from 90+ options and labeling who said what. You get a timestamped transcript synced to the audio, so you can click any word and hear that exact moment, plus a per-word confidence view that flags anything worth double-checking. From there you can edit the text, hide filler words, and export to TXT, DOCX, SRT, or WebVTT. The first 10 minutes of every file are transcribed free on a free account, no credit card required; beyond that it is $3 per audio hour, billed per started hour, with a $3 minimum per file. This works on recordings, not live dictation: record first on any device, then upload the file.
No. Dictation types out what you say while you speak: your phone keyboard's microphone key and Google Docs voice typing do that, and they are free. But dictation cannot open an existing recording, tell speakers apart, or add timestamps. FileToText is speech to text for recordings: upload a file after the fact and get a speaker-labeled, timestamped transcript you can verify against the audio.
In practice they mean the same thing: turning spoken words into written text. People searching voice to text usually mean short personal recordings, a voice memo or a voicemail, while speech to text covers everything from a quick memo to a 20-hour recording. FileToText handles both the same way: upload the file, get the transcript.
The first 10 minutes of every file are free, on every account, with no credit card. That covers most voice memos and voicemails in full. Longer files cost $3 per audio hour, billed per started hour, with a $3 minimum per file, or a subscription from $10/month (600 minutes) if you transcribe regularly. Not satisfied? Full refund, no questions asked.
95%+ on clear audio. Background noise, strong accents, distant microphones, and people talking over each other pull that down. The per-word confidence view highlights the words the engine was least sure about, and clicking any word replays that moment so you can check it yourself. See the live monthly benchmark →
More than 90 languages, detected automatically. You do not set the language by hand: upload the file and the engine identifies what is spoken automatically.

Start transcribing here