How to Transcribe Audio to Text (Free and Paid Options)

The fastest way to transcribe audio to text is to upload the recording to a free AI transcriber and copy the result, which takes about a minute. If you'd rather stay inside apps you already have, Microsoft Word, iPhone Voice Memos, Apple Notes and the Pixel Recorder all transcribe for free (or with a subscription you may already pay for). For private or very long files, you can run OpenAI's Whisper on your own computer.

1Transcribe audio to text free online

If you already have a recording (a voice memo, an MP3, a WhatsApp voice note, a Zoom recording, a lecture), an online AI transcriber is the quickest option. Our free Audio & Video to Text tool needs no account:

1. Open urltotext.com/audio-to-text.
2. Click Choose an audio or video file, or drag the file onto the box. The upload starts right away.
3. Wait while it transcribes. Most files take 10 to 60 seconds.
4. Click Copy, or Download .txt to save the transcript as a text file.

What it handles:

• Files: MP3, M4A, WAV, AAC, OGG, OPUS, FLAC, AMR, MP4, MOV, MKV, WebM and more, up to 95 MB, including iPhone voice memos.
• Length: up to 15 minutes per recording without an account, or up to 60 minutes with a free account.
• Runs: 5 an hour without an account, 30 an hour signed in. It's free either way.
• Languages: detected automatically, including English, Spanish, Portuguese, French, German, Hindi, Arabic, Japanese and Chinese.
• Links too: switch to the Paste a link tab to transcribe a YouTube, TikTok, Instagram, Vimeo, Loom or podcast link instead of a file.

It runs on Whisper Large v3 Turbo, one of the most accurate open speech models, and your file is deleted as soon as it's transcribed.

Limitations: the transcript is plain text, without speaker labels or timestamps. For timestamps, use the Subtitle Generator, which gives you the same transcript as timed SRT or VTT captions. Want the key points rather than every word? The Video Summarizer transcribes and summarizes in one step, and Translate Video transcribes and translates into 30+ languages.

Tip: if a video is over 95 MB, export just the audio as M4A or MP3 first. It's far smaller.

2Microsoft Word for the web: Transcribe

If you pay for Microsoft 365, Word has a built-in transcriber that separates speakers and adds timestamps.

1. Go to office.com in Microsoft Edge or Chrome and sign in. Transcribe doesn't work in Safari or Firefox.
2. Open a new or existing Word document.
3. On the Home tab, open the Dictate dropdown (the arrow next to the microphone) and choose Transcribe.
4. In the Transcribe pane, click Upload audio and pick your file. To record a conversation live instead, click Start recording.
5. Keep the pane open while it works. It can take about as long as the recording itself.
6. Review the transcript in the pane, then click Add all to document (or Add section to document for one part).

Speaker labels: Word marks each voice as Speaker 1, Speaker 2 and so on. To rename one, hover over a section, click Edit transcript section, change the name, tick Change all Speaker 1 to rename every instance, and click Confirm.

Limits:

• Uploads must be .wav, .mp4, .m4a or .mp3.
• A Microsoft 365 subscription includes 300 minutes of uploaded audio a month. A Microsoft 365 Copilot license raises that to 30,000 minutes.
• Each document holds one transcript. Starting New transcription replaces it (sections you already added stay).
• Your audio is saved to a Transcribed Files folder in your OneDrive.

The free Word web app without a subscription doesn't include Transcribe.

3Google Docs voice typing

Google Docs can't transcribe an uploaded file, but its free voice typing writes down whatever your microphone hears. It's great for dictation and can handle a short clip played aloud.

1. Open a document in Chrome, Edge or Safari.
2. Click Tools›Voice typing, or press Ctrl + Shift + S (Cmd + Shift + S on a Mac).
3. Pick your language from the dropdown above the microphone.
4. Click the microphone (it turns red) and allow microphone access if asked.
5. Speak, or play your recording. Click the microphone again to stop.

In English, say "period", "comma", "question mark" or "new paragraph" to punctuate, and "voice commands list" to see every command.

The speaker trick for recordings: play the file through your speakers in a quiet room with the mic listening. It works, but echo and any noise in the room cost accuracy. A virtual audio device (Stereo Mix on some Windows PCs, BlackHole on a Mac) that feeds the computer's sound straight into the mic input is cleaner.

Limitations:

• It's real time: a 40-minute file takes 40 minutes, with the tab open and in focus. It also stops after a long silence.
• No speaker labels or timestamps. For a file you already have, an upload transcriber is faster and more accurate.

4iPhone Voice Memos and Apple Notes

On an iPhone 12 or later, Voice Memos and Notes transcribe recordings for free, on the phone. Transcription works in English, Spanish, Portuguese, Italian, French, German, Japanese, Korean and Chinese (Simplified and Traditional), though not in every country or region.

Voice Memos

1. Open Voice Memos (it's in the Utilities folder if it's not on your Home Screen).
2. Tap the red Record button. To watch the words appear live, swipe up from the top of the waveform and tap the Transcription button (a speech bubble).
3. Tap Done to save.
4. To copy the text later, tap the recording, tap the More button (three dots), then choose Copy Transcript for all of it, or View Transcript to select just part and tap Copy.

Search in Voice Memos looks inside transcripts too, so you can find a recording by something said in it.

Apple Notes

1. In a note, tap the Attachments button (the paperclip), then Record Audio.
2. Tap Record. Tap the Transcription button to see the live transcript.
3. When you're finished, tap Pause, then Done.
4. Tap the recording, tap the Transcription button, then tap More and choose Add Transcript to Note or Copy Transcript. Find in Transcript searches it.

Phone and FaceTime calls you record (iOS 18.1 or later) are transcribed in Notes too, and with Apple Intelligence you can tap Summary for a recap.

Limitations: no speaker labels, and no easy way to transcribe an audio file you got from someone else. For that, save it to Files and upload it to Audio & Video to Text. Voice memos are M4A files and upload as they are; tap the recording, then the Share button, then Save to Files.

5Android: Pixel Recorder and Live Transcribe

Google Recorder (Pixel phones and Pixel Tablet)

The Recorder app on Pixel 3 and later transcribes while you record, on the device, without an internet connection.

1. Open Recorder and tap Record.
2. While it records, tap Transcript to switch from the waveform to the live text.
3. Tap Pause, then Save. You can record up to 18 hours in one recording.
4. To get the text out, touch and hold the recording, tap Share, then Transcript (.txt). Or open the recording, tap More›Share›Share transcript to Google Docs.

On a Pixel 6 or later, turn on speaker labels under your profile icon, Recorder settings›Speaker labels (US English only). Fix a wrong word by touching and holding it and tapping Edit word. Pixel 6 and later transcribe English, Mandarin Chinese, French, German, Hindi, Italian, Japanese and Spanish; Pixel 3 to 5a handle English, French, German and Japanese.

Live Transcribe (most Android phones)

Google's free Live Transcribe & Notification app captions conversations in real time. On a Pixel, open Settings›Accessibility›Live Transcribe; on other phones, install it from Google Play. It's for reading along, not record keeping: transcripts aren't saved unless you turn saving on in its settings, and then only for three days.

6Otter.ai and other transcription apps

If you transcribe meetings or interviews every week, a dedicated app adds speaker labels, editing, search and a bot that joins your calls.

Otter.ai is the best known. Its plans as of October 2026:

• Basic (free): 300 minutes a month, up to 30 minutes per conversation, and only 3 file imports for the life of the account.
• Pro: $16.99 a month, or $8.33 a month billed yearly. 1,200 minutes a month, 90 minutes per conversation, 10 imports a month.
• Business: $30 a month, or $19.99 billed yearly. Up to 4 hours per conversation and unlimited imports (6,000 minutes a month for imported files).

To transcribe a file, sign in at otter.ai, click Import and choose it. With only three free imports, the free plan suits live meetings more than a backlog of files.

Alternatives: TurboScribe (free for 3 files a day, up to 30 minutes each), Descript (edit the audio by editing its transcript) and Rev (paid human transcription when it must be near perfect). Prices change often, so check before you sign up.

7Run Whisper on your own computer

Whisper is OpenAI's open-source speech recognition model, free to run on your own machine. Nothing leaves your computer, there's no length limit, and you get timestamps and subtitle files. You'll need a terminal.

OpenAI Whisper (Python)

1. Install Python 3 and ffmpeg (Mac: brew install ffmpeg; Ubuntu: sudo apt install ffmpeg; Windows: winget install Gyan.FFmpeg or choco install ffmpeg).
2. Install Whisper:

pip install -U openai-whisper

3. Transcribe a file:

whisper interview.mp3 --model turbo

Whisper writes .txt, .srt, .vtt, .tsv and .json files in the current folder. Add --output_format txt for just the text and --language English to skip detection. turbo wants a GPU with about 6 GB of memory; on a laptop without one, --model small is faster and a little less accurate.

faster-whisper (quicker, less memory)

faster-whisper runs the same models up to four times faster. Run pip install faster-whisper, then:

from faster_whisper import WhisperModel

model = WhisperModel("large-v3", device="cpu", compute_type="int8")
segments, info = model.transcribe("interview.mp3", vad_filter=True)
for s in segments:
    print(f"[{s.start:.0f}s] {s.text}")

whisper.cpp (no Python, great on Macs)

whisper.cpp is a C++ port that runs well on Apple Silicon and ordinary CPUs:

git clone https://github.com/ggml-org/whisper.cpp
cd whisper.cpp
sh ./models/download-ggml-model.sh base.en
cmake -B build && cmake --build build -j --config Release
ffmpeg -i interview.mp3 -ar 16000 -ac 1 -c:a pcm_s16le interview.wav
./build/bin/whisper-cli -m models/ggml-base.en.bin -f interview.wav -otxt

whisper-cli needs 16-bit WAV input, which the ffmpeg line makes. Use -osrt for subtitles.

Limitations: none of these label speakers (WhisperX adds that), and big models are slow without a GPU. For a few files, our online tool runs the same Whisper Large v3 Turbo model for you.

8Transcribe Zoom, Google Meet and Teams meetings

For a call, switch on the meeting app's own transcript. Tell people first; these apps notify everyone anyway.

Zoom: audio transcripts come with cloud recording on paid plans (Pro, Business, Education or Enterprise). In the Zoom web portal, go to Settings›Recording, turn on Cloud recording and tick Create audio transcript. Record the meeting to the cloud; afterward, find the transcript under Recordings & Transcripts›Cloud recordings, where it shows beside the video and downloads as a .vtt file. On the free plan, record to your computer instead and upload the audio file Zoom saves (an M4A, or the MP4) to Audio & Video to Text.

Google Meet: on a computer, click Meeting tools›Transcribe›Start transcription, then Start. The transcript is saved as a Google Doc in the organizer's Drive, in the Google Meet folder. It needs a Workspace plan with transcripts (Business Standard and up, Workspace Individual, some Education plans) and works in eight languages, including English, Spanish and French.

Microsoft Teams: in the meeting, click More actions›Record and transcribe›Start transcription and confirm the spoken language. Starting a recording also starts the transcript. Afterward, open the meeting's Recap tab (or the meeting chat) to read the transcript and download it as .docx or .vtt. This needs a work or school account where your admin allows transcription.

9Transcribe audio by hand

When you need every "um" or exact quotes, you'll type it yourself. Plan on about four hours per hour of clear audio. oTranscribe is a free web app that puts the player and a text editor on one page:

1. Open oTranscribe and click Start transcribing.
2. Choose your audio or video file (or paste a YouTube link).
3. Type in the editor while you control playback from the keyboard.

Its default shortcuts:

• Esc: play or pause
• F1 / F2: jump back / forward a few seconds (or Ctrl + 1 / Ctrl + 2; Cmd on a Mac)
• F3 / F4: slow down / speed up
• Ctrl + J: insert a clickable timestamp

Faster hybrid: let AI write the first draft, then listen through in oTranscribe at 1.25x speed and fix mistakes.

Speaker labels, timestamps and accuracy tips

If you need speaker labels, use Word's Transcribe, the Pixel Recorder (US English), Otter, or a meeting app's transcript. Plain Whisper, Voice Memos and Google Docs don't separate speakers.

If you need timestamps, use the Subtitle Generator (SRT or VTT), Whisper's .srt output, Word's transcript, or oTranscribe's Ctrl + J timestamps.

To get a more accurate transcript:

• Record close to the speaker. A phone a foot from someone beats a laptop across the room. A cheap clip-on mic helps a lot for interviews.
• Cut background noise (fans, music, echoey rooms). AI copes with accents far better than with clatter.
• Avoid cross-talk. People talking over each other is where every tool fails most.
• Upload the original file, not a low-bitrate copy. M4A, MP3 at 128 kbps or more, WAV and FLAC are all fine.
• Check names and numbers first. Proper nouns, jargon and figures are the most common mistakes.

Privacy: where does your audio go?

• On your device: Voice Memos, Notes, the Pixel Recorder and Whisper. Best for anything sensitive.
• Uploaded, then deleted: Audio & Video to Text deletes your file as soon as it's transcribed.
• Kept in the cloud: Word (OneDrive), Otter, Zoom, Meet and Teams store recordings until you delete them. Check your organization's rules for client, patient or legal audio.
• Consent: in many US states and countries, everyone must agree to be recorded. Ask first.

Frequently asked questions

How can I transcribe an audio file for free?

Upload it to Audio & Video to Text and click Copy when it's done. It's free for files up to 15 minutes without an account, or 60 minutes with a free account. On an iPhone, Voice Memos transcribes your own recordings for free. For unlimited free transcription, run Whisper on your computer.

How do I convert audio to text in Word?

In Word for the web (Edge or Chrome), go to Home›Dictate›Transcribe, click Upload audio, and choose a .wav, .mp4, .m4a or .mp3 file. Then click Add all to document. It needs a Microsoft 365 subscription and includes 300 minutes of uploads a month.

Can Google Docs transcribe an audio file?

Not directly. Google Docs voice typing only listens to your microphone, so you'd have to play the file aloud and let it type in real time. An upload transcriber is faster and more accurate for recordings.

How do I transcribe a voice memo on iPhone?

On an iPhone 12 or later, open Voice Memos, tap the recording, tap the More button and choose Copy Transcript. Paste it into Notes, Mail or any app. On older iPhones, share the memo to Files and upload it to an online transcriber.

What is the most accurate way to transcribe audio?

On clear audio, Whisper-based tools, Otter and Word are all very accurate. For near-perfect results, correct an AI draft yourself or pay for human transcription. Recording quality matters more than the tool.

Which method should you use?

• One file, right now: Audio & Video to Text. Free, no account, about a minute.
• Recording on your phone: Voice Memos on iPhone, Recorder on a Pixel.
• Already pay for Microsoft 365: Word's Transcribe, for speaker labels.
• Weekly meetings: your meeting app's transcript, or Otter.
• Sensitive or very long recordings: Whisper on your own computer.
• Every word must be right: AI first draft, then a careful pass in oTranscribe.