The best free audio transcription software in 2026

Short answer:

• One recording turned into text, free, nothing uploaded, nothing to install: the free Audio to text tool on this site. It runs Whisper in your browser and saves TXT, Word, SRT or VTT.
• The most capable free app on a Mac: MacWhisper, which runs Whisper models on your Mac.
• Free and open source, and you are fine with a command line: OpenAI's Whisper, or Buzz for a window to click.
• Meetings, with speaker names and a summary: Otter's free plan, 300 minutes a month, 30 minutes per conversation.
• Typing it yourself while you listen: oTranscribe.

Not for you if you need a certified human transcript for a court or a medical record: none of these free tools are that, and a paid human service is the honest answer.

How we checked these tools

We make MotionShine and the Audio to text tool on this site, so weigh what we say about it the way you weigh any vendor's own page. Every fact about the other tools comes from the maker's own website, pricing page, help center or GitHub page, checked on 7 November 2026. Where the maker does not say, the table says "not stated". We did not measure accuracy, and we do not quote accuracy numbers: how well any of these tools does depends far more on your recording than on the tool.

Which job do you have?

All 9 at a glance

Facts from each maker's own page, checked on 7 November 2026.

The best free audio transcription software, one by one

1. MotionShine Audio to text (our tool)

Price: free, no sign-up, no minute limit, no watermark.

What it does: drop in an MP3, M4A, WAV, OGG or FLAC file, or a MOV, MP4 or WebM video, and it writes the words with the time of each line. It detects the spoken language or you pick it. The transcript opens as lines with a player above, so you click a time to hear that line and click the line to fix a word, with find and replace for a name that is wrong everywhere. It saves plain text, text with times, Word, SRT or VTT, and it can translate the transcript. It runs Whisper base, OpenAI's open speech recognition model, in your browser; the model downloads once, about 95 MB, and your recording is never uploaded.

Good for: interviews, lectures, voice memos and video subtitles, when you would rather not send the file to a server or install an app.

Weak point: Whisper base is one of the small versions of Whisper, picked so it fits in a web page. The larger versions, which MacWhisper and the Whisper command line can run, are more accurate. It is also slower than a cloud service on many computers, and it does not say who is speaking; for labels such as Speaker 1 and Speaker 2 there is a separate speaker identification tool.

2. Apple Notes (built into the Mac)

Price: free.

What it does: Apple's guide says you open a note, click the Attachment button, choose Record Audio, and click the Transcription button to see the live transcript. Afterwards you can copy the transcript or add it to the note. It needs a Mac with M1 or later, and works in English (several regions), Spanish, Portuguese, Italian, French, German, Japanese, Korean and Chinese.

Good for: recording a lecture or a conversation on your Mac and getting the text at the same time.

Weak point: Apple's page describes recording in Notes, not importing an existing file, so it does not help with a recording you already have. There is no SRT export.

3. MacWhisper

Price: free to use; Pro is €64 once, with lifetime updates, per its site.

What it does: a Mac app that transcribes files with Whisper models. Its site says local models process "without data ever leaving your Mac", and lists speaker labels, filler word removal, batch transcription and export to subtitles and documents. It can also use cloud models and outside AI providers.

Good for: people who transcribe often on a Mac and want bigger models and batch work.

Weak point: the main page does not say which features are free and which are Pro. Your data stays on the Mac only when you use the local models.

4. OpenAI Whisper (command line)

Price: free; code and model weights under the MIT licence.

What it does: the open source speech recognition model itself. Its GitHub page lists model sizes from tiny (39 million parameters) to large (1,550 million) and turbo, and commands such as whisper audio.mp3 --model turbo. It also translates speech into English, and its source code writes TXT, SRT, VTT, TSV and JSON files.

Good for: developers and anyone who wants the larger models for free and can install Python and ffmpeg.

Weak point: there is no app window. The bigger models need a capable computer; the page lists about 10 GB of video memory for large.

5. Buzz

Price: free and open source on GitHub.

What it does: its site describes "offline audio transcription and translation" powered by Whisper. It imports audio and video files, exports CSV, SRT, TXT and VTT, and transcribes live from your microphone in more than 90 languages.

Good for: a free desktop app with a window to click, without a monthly limit.

Weak point: Buzz Classic is listed for Windows, Linux and Intel Macs; its site also shows a separate Buzz for macOS app, whose price is not stated there. The site warns that Whisper is resource-intensive.

6. Otter

Price: Basic is free; Pro is $16.99 per user a month, or $8.33 a month billed yearly.

What it does: Otter's pricing page lists the free plan as 300 transcription minutes a month per user, conversations up to 30 minutes each, and 3 file imports in total for the life of the account.

Good for: live meetings, where Otter joins and transcribes as it happens.

Weak point: for files, the 3 lifetime imports end the free plan quickly. Your audio is processed in the cloud.

7. Notta

Price: free plan; Pro is $8.17 a month billed yearly ($97.99 a year).

What it does: Notta's pricing page lists the free plan as 120 transcription minutes a month, up to 3 minutes per conversation, and 50 file uploads a month.

Good for: trying a cloud transcription service on short clips.

Weak point: 3 minutes per conversation rules out a whole interview or lecture on the free plan.

8. Descript

Price: Free; Hobbyist is $24 a month, or $16 a month billed yearly.

What it does: an audio and video editor where you edit by changing the transcript. The free plan has 60 media minutes a month and transcription in 25 languages, per its pricing page.

Good for: podcasters and video makers who want to cut the recording by deleting words.

Weak point: free video exports are 720p with a watermark, and the 60 minutes count whether or not you transcribe.

9. oTranscribe

Price: free and open source (MIT).

What it does: not automatic transcription. It is a web page that plays your recording next to a text box, with keys to pause, rewind and add a timestamp while you type. Its site says your audio file and transcript never leave your computer, and it exports Markdown, plain text or Google Docs.

Good for: careful transcripts you write yourself, or checking an automatic one.

Weak point: you do all the typing, and it works on desktop computers only.

Free, free plan or trial: what the word means

• No allowance at all: Audio to text, Apple Notes, Whisper, Buzz and oTranscribe. MacWhisper has a free version and a paid Pro.
• A monthly allowance: Otter (300 minutes), Notta (120 minutes), Descript (60 media minutes).
• Not free: Microsoft Word's Transcribe needs a Microsoft 365 subscription, and Microsoft's page says it is available in Word for Microsoft 365 on Windows, with up to 300 minutes of uploaded audio a month.

Does your recording get uploaded?

Cloud services (Otter, Notta, Descript) send your audio to their servers. That is how they offer summaries and sharing. For an interview with a source, a patient or a client, a tool that runs on your own computer keeps the recording with you: Audio to text, MacWhisper with local models, Whisper, Buzz and oTranscribe all work that way.

How to get a cleaner transcript with any tool

• Record close to the mouth. A phone held near the speaker beats a laptop across the room.
• Clean the sound first. Hum and traffic can lead to words nobody said in quiet gaps. Run the file through the background noise remover before you transcribe, and see how to remove background noise from a video.
• Set the language if the tool lets you, instead of leaving it to guess on a mixed recording.
• Check names and numbers against the recording before you quote them. Every automatic tool gets some wrong.

If you are recording the meeting yourself, how to record a meeting on Mac and the best audio recorders for Mac cover that part.

Keep a paid service if

• You need speaker labels, summaries and search across hundreds of meetings, shared with a team.
• You transcribe many hours a week and want it done fast on someone else's computers.
• You need a human-checked transcript you can stand behind.

Which one should you pick?

• One file, private, today: Audio to text. For a video's sound track, video to text does the same.
• Lots of files on a Mac: MacWhisper.
• Free and open source: Whisper on the command line, or Buzz.
• Meetings: Otter's free plan.
• Subtitles for a screen recording: MotionShine, by the maker of this site, makes captions on your Mac in its editor. Free exports go up to 720p at 30 fps with no watermark.

FAQ

What is the best free audio transcription software?

It depends on the job. For one file with nothing uploaded and nothing to install, the Audio to text tool on this site. For heavy use on a Mac, MacWhisper. For meetings, Otter's free plan. For free and open source, Whisper or Buzz.

Is there a free transcription app with no time limit?

Yes. Audio to text, Whisper, Buzz and oTranscribe have no minute limit, because they run on your own computer. Cloud services set monthly limits: Otter 300 minutes, Notta 120, Descript 60 media minutes, per their pages on 7 November 2026.

Can I transcribe audio to text without uploading it?

Yes. Audio to text runs Whisper in your browser tab, MacWhisper runs local models on your Mac, and Whisper and Buzz run on your own computer. In each case the recording stays with you.

How do I transcribe a voice memo on a Mac?

Voice Memos saves M4A files. Drop the file into Audio to text, or open it in MacWhisper. Apple Notes transcribes while it records, but Apple's page does not describe importing an existing memo.

Can free transcription software tell who is speaking?

Some can. MacWhisper's site lists speaker labels, and Otter is built for meetings. The Audio to text tool does not; use the speaker identification tool for that.

Can I get subtitles from a transcription?

Yes. Audio to text, Buzz and MacWhisper export SRT or VTT files with timings, which YouTube and most video editors read.