How to transcribe audio to text: 4 ways, compared
October 30, 2026 · 6 min read
There are four realistic ways to turn a recording into text: upload it to an AI transcription tool, use Microsoft Word’s Transcribe feature, try Google Docs voice typing, or type it out yourself. For most recordings an AI tool is the fastest and cheapest: an hour of audio comes back as text in minutes, and you spend your time checking it rather than typing it.
Here is how each one works, what it costs, and when it’s the right choice.
The short version
| Method | Works with audio files? | Time for 1 hour of audio | Cost |
|---|---|---|---|
| AI transcription tool | Yes, most formats | A few minutes, plus checking | Free tiers; paid from a few dollars |
| Microsoft Word Transcribe | Yes (MP3, M4A, WAV, MP4) | A few minutes | Needs Microsoft 365; monthly upload limit |
| Google Docs voice typing | No, microphone only | At least 1 hour, with gaps | Free |
| Typing it yourself | Yes | Often 3–4 hours or more | Free, but your time |
1. Use an AI transcription tool
AI transcription tools run your recording through a speech recognition model and give you back text, usually with timestamps. Modern models handle accents, background noise and long recordings far better than the dictation features of a few years ago.
How to do it with TranscribeNest (most tools work the same way):
- Go to the upload box on the home page and drop in your file. MP3, M4A, WAV, OGG, WhatsApp voice notes and video files such as MP4 and MOV all work. If it’s a video, only the audio track is extracted and uploaded, so even large files go up quickly.
- Leave the language on auto-detect, or pick it yourself if the recording has a strong accent or mixes languages.
- Wait a few minutes. Long recordings are split into parts and transcribed in parallel.
- Read the transcript on the page. Click any line to hear that moment, and search for words you want to check.
- Copy the text or download it as a Word document, plain text, or SRT and VTT subtitles.
Your first file doesn’t need an account, and after that you get 3 free files a day of up to 30 minutes each. Longer recordings have their first 30 minutes transcribed for free; the rest needs bought hours or a subscription.
What to check afterwards. Speech recognition is most often wrong on names, numbers, technical terms and words that sound alike. Search for the names that matter, listen to the lines with numbers, and fix them. On a clear recording that’s usually a few minutes of work per hour of audio.
Best for: almost everything, especially interviews, lectures, meetings, podcasts and voice memos longer than a few minutes.
2. Use Microsoft Word’s Transcribe feature
If you have a Microsoft 365 subscription, Word can transcribe recordings itself. In Word for the web, go to Home, open the Dictate menu and choose Transcribe, then upload your file. Microsoft lists .wav, .mp4, .m4a and .mp3 files and a limit of 300 minutes of uploaded audio per month for subscribers; check Microsoft’s help page for “Transcribe in Word” for the current details.
The transcript appears in a side panel, and you can add it to your document with or without speaker names and timestamps.
Best for: people who already pay for Microsoft 365 and transcribe less than five hours a month. If you don’t have a subscription, an AI tool can give you a Word document anyway: see how to transcribe audio to a Word document.
3. Use Google Docs voice typing (with caveats)
Google Docs has voice typing (Tools, then Voice typing, in Chrome on a computer). It types what it hears through your microphone, live. It can’t open an audio file.
Some people play a recording through their speakers next to the microphone. It works poorly: it only runs in real time, the sound loses quality on the way, typing stops at pauses, and you have to sit next to it for the whole recording. It’s fine for dictating your own notes, not for transcribing recordings. More on this in why Google Docs can’t transcribe audio files, and what to do instead.
Best for: dictating new text, not transcribing existing recordings.
4. Type it yourself
Manual transcription is still the right choice when the audio is very poor, when every word carries legal weight, or when you need a strict verbatim transcript with every pause and “um” marked.
A setup that makes it bearable:
- A player built for transcription. Free web tools such as oTranscribe put the audio player and a text editor on one page, with keyboard shortcuts to pause and jump back a few seconds without touching the mouse.
- Slower playback. At 0.7× or 0.8× speed you can type along with fewer pauses.
- A rhythm: listen to a short phrase, pause, type, jump back two seconds, continue.
- Time markers every few minutes, so you can find your place later.
Expect it to take time. A commonly quoted rule of thumb is 3–4 hours of typing for every hour of clear audio, and more for poor audio or several speakers.
A good compromise is to let an AI tool make the first draft and correct it by hand while listening. That turns hours of typing into a careful read-through.
Best for: very poor audio, legal or research work that needs strict verbatim, or very short clips.
How to get a more accurate transcript
Whatever method you use, the recording matters more than anything else:
- Get the microphone close. A phone on the table in the middle of a meeting picks up the room; a microphone near the speaker picks up the voice.
- Cut background noise. Music, air conditioning and cafés hurt accuracy more than accents do.
- One person at a time. Overlapping voices are where every transcription method struggles.
- Choose the language if auto-detect gets it wrong, for example on a short clip or a recording that starts in another language.
After transcribing: turn it into a usable document
A raw transcript is a starting point. Depending on what it’s for you may want to mark who is speaking, add a header with the date and participants, keep timestamps next to quotes, or remove filler words. Our guide on how to write a transcript covers formats, speaker labels and timestamps with examples.
FAQ
What’s the fastest way to convert audio to text? Upload it to an AI transcription tool. An hour of audio usually comes back in a few minutes.
Can I transcribe audio for free? Yes. TranscribeNest’s first file needs no account, then you get 3 free files a day of up to 30 minutes each. Typing it yourself is free too, but takes hours.
Can I transcribe a video the same way? Yes. Upload the video file; the audio is extracted in your browser and transcribed. You can also download the result as SRT subtitles for the video. See video to text.
Does it work in other languages? AI tools handle many languages. TranscribeNest supports 90+ and detects the language automatically.