How to Transcribe Audio for Free: A Complete Audio-to-Text Guide

To transcribe audio for free, upload your recording to a free speech-to-text tool, pick the spoken language (or leave it on auto-detect), and let the software turn the speech into timestamped text you can copy or download. With WaveToWord, you sign in with a free account, open the Audio to text tool, drop in an MP3, WAV, M4A or other common audio file, and get the transcript back as plain text or as SRT and VTT subtitle files. Your phone and computer may also have built-in options that cost nothing. The rest of this guide shows you how each method works, how to get the cleanest possible result, and which option suits which kind of recording.

Key takeaways

• The fastest free route is an online audio-to-text converter: upload, choose a language, download the transcript.
• WaveToWord's Audio to text tool is free with a free account. It accepts MP3, WAV, M4A, AAC, OGG, OPUS, FLAC, WMA and WebM files and handles recordings up to 15 minutes per run.
• Built-in options exist on iPhone (Voice Memos), Pixel phones (Recorder) and in Microsoft Word for the web, but each has device, language or subscription limits.
• Audio quality matters more than the software. A close microphone, a quiet room and one person talking at a time give you far fewer errors.
• Always proofread names, numbers and quotes before you publish or rely on a machine transcript.

How do I transcribe audio to text? The short version

Every automatic transcription method follows the same three-step pattern, whether you use a website, a phone app or a word processor:

1. Get the audio into the tool. Upload a file, record directly, or point the tool at an existing recording.
2. Tell it what language is spoken. Most tools offer auto-detect, but choosing the language yourself usually improves accuracy, especially for short clips or strong accents.
3. Review and export. Read through the text, fix mistakes, and save it in the format you need: plain text for notes, a document for sharing, or a subtitle file for video.

What separates good results from frustrating ones is mostly what happens before step one (how the audio was recorded) and after step three (how carefully you check the output). We cover both below.

How to transcribe audio for free with WaveToWord

WaveToWord is a free web app for turning speech into text. There is nothing to pay: you create a free account, sign in, and use the tools. Here is the exact workflow for an audio file.

1. Open the tool. Go to Audio to text and sign in with your free account.
2. Add your file. Drag the recording onto the page or browse for it. Supported formats are MP3, WAV, M4A, AAC, OGG, OGA, OPUS, FLAC, WMA and WebM audio, up to 500 MB per file for signed-in users.
3. Choose the language. Keep auto-detect or pick the spoken language from the list of several dozen supported languages. If you know the language, select it.
4. Start the run. Processing happens on our servers while you wait on the page. A short voice note usually finishes quickly; longer recordings take a bit more time.
5. Read the result. The transcript appears as timestamped segments. When a line looks wrong, the timestamp tells you exactly where to listen in your original recording.
6. Copy or download. Use Copy text to paste it anywhere, or download it as TXT (plain text), SRT or VTT (subtitle files with timings).

Limits you should know about

Being upfront about limits saves you a wasted upload:

If your recording runs longer than 15 minutes, use the Audio cutter to split it into sections and transcribe each one. If the file is huge because it is an uncompressed WAV, converting it to MP3 or M4A with the Audio converter shrinks it dramatically without hurting speech recognition.

Keeping your transcripts in a library

The free tools are built for quick, one-off jobs. If you transcribe regularly, your free account also gives you a Files library. Recordings you upload there are not limited to 15 minutes (a monthly fair-use allowance applies instead), and you get a transcript you can edit line by line while the audio plays, with speakers you can rename and reassign, folders, and exports such as DOCX, PDF, SRT, VTT and plain text. That is the better choice for long interviews, research recordings or anything you will come back to.

Free audio to text converter options compared

There is no single best free tool for every recording. The table below compares the main free routes so you can pick the one that fits your file, device and privacy needs.

A few practical notes on the built-in options:

• Phone recorders (iPhone Voice Memos, Pixel Recorder, Samsung Voice Recorder). Recent phones can transcribe recordings made in their own recorder app, in a limited set of languages. They are convenient for memos, but they do not import arbitrary files or export subtitle timings. Our guide on turning a voice memo into text has the exact steps for each phone.
• Microsoft Word for the web. Subscribers can upload an audio file through the Dictate menu's Transcribe option and get a speaker-separated transcript inside a document. It is handy if you already pay for Microsoft 365, but it is not free on its own and it has a monthly upload allowance.

How to transcribe an MP3 audio file to text step by step

MP3 is the most common format people want to convert, so here is a fuller walk-through that covers preparation, the transcription itself, and cleanup. The same steps work for M4A recordings from phones and WAV files from recorders.

Step 1: Check the file before you upload

Open the MP3 and listen to the first 30 seconds and a random minute in the middle. You are checking three things:

• Is speech the loudest thing? If music or room noise competes with the voice, expect more errors.
• Is there long silence at the start? Trim dead air at the beginning and end with an audio cutter. It saves processing time and avoids odd output at the edges.
• How long is it? If it is longer than 15 minutes, split it into parts now.

Step 2: Pick the right language setting

Auto-detect works well when a recording starts with clear speech in one language. It can stumble when the first few seconds are music, a greeting in another language, or a very short clip. If you know the language, select it manually. For recordings where people switch between two languages, choose the dominant one and expect to correct the other.

Step 3: Transcribe

Upload the MP3 to Audio to text, choose the language, and start the run. Stay on the page until it completes.

Step 4: Proofread with the audio

Do not just skim the text. Play the audio and follow along, paying special attention to:

• Proper nouns: people, companies, product names and places are the most common errors.
• Numbers and units: "fifteen" versus "fifty," dates, prices and percentages.
• Homophones: "their/there," "affect/effect," and industry jargon that sounds like everyday words.
• Overlapping speech: where two people talked at once, words are often dropped or merged.

Step 5: Export in the right format

• Choose TXT for notes, quotes, blog drafts or pasting into a document.
• Choose SRT if you will upload captions to a video platform or editor; it is the most widely supported subtitle format.
• Choose VTT for web video players that use the WebVTT standard.

If you are unsure which subtitle format to use, our explainer on SRT vs VTT subtitle files breaks down the differences.

How accurate is free online transcription of audio to text?

Modern speech recognition is very good on clean audio and noticeably worse on messy audio. The software is rarely the bottleneck; the recording is. In our experience working with interviews, lectures and voice notes, these are the factors that move accuracy the most, in rough order of impact.

Microphone distance

The single biggest improvement you can make is putting the microphone close to the speaker's mouth. A phone lying on a conference table two meters away picks up as much room echo as voice. A phone held 15 to 30 centimeters from the speaker, or a clip-on lavalier mic, captures speech that transcribes far more cleanly. For interviews, place the recorder between the two people and closer to whoever speaks more quietly.

Background noise

Fans, air conditioning, traffic, cafe chatter and music all compete with speech. Steady noise like a fan is easier for software to ignore than unpredictable noise like clattering dishes or a TV. If you can choose the room, pick a small space with soft furnishings; carpets, curtains and sofas reduce echo.

Overlapping speakers

When people interrupt each other, transcription engines struggle to separate the voices. In meetings, a quick reminder to let people finish their sentences helps both the transcript and the meeting. If speaker labels are important, ask each person to introduce themselves at the start so you can tell voices apart when proofreading.

Accents, pace and jargon

Clear, moderately paced speech transcribes best. Strong regional accents, very fast talkers and specialist vocabulary (medical terms, legal Latin, product codes) increase errors. Selecting the correct language variant and doing a focused proofread on technical terms is the fix.

File quality and format

Speech recognition does not need audiophile quality. A 128 kbps MP3 or a standard phone M4A is plenty. What hurts is heavy compression at very low bitrates, clipping (audio recorded so loud that it distorts), and recordings that were already converted several times. If you have the original file, use it rather than a copy that was re-shared through messaging apps. For more on formats, see WAV vs MP3.

Common transcription mistakes and how to fix them

Most problems people run into fall into a handful of patterns. Here is what causes each and what to do about it.

The transcript is in the wrong language or full of gibberish. Auto-detect guessed wrong, usually because the clip starts with music or silence. Re-run with the language selected manually, and trim the intro.

Whole sentences are missing. Look at where the gaps are. If they line up with quiet speakers or people far from the microphone, the audio did not capture them clearly enough. Boosting the volume of the file will not recover detail that was never recorded, but trimming noisy sections and re-running can help.

Words repeat or the text drifts away from what was said. This can happen with long stretches of silence, music or noise, where an engine tries to find speech that is not there. Cut out non-speech sections before transcribing.

The upload is rejected. Check the format and size. Video files belong in Video to text rather than the audio tool, and anything over 15 minutes needs splitting. If you have an unusual format, convert it to MP3 first.

Timestamps in the subtitle file are slightly off. Machine timings are close but not frame-perfect. For video captions, open the SRT in a subtitle editor, shift the timings, and fix line breaks before publishing.

Names are spelled several different ways. Use find-and-replace in your text editor once you have confirmed the correct spelling. In the WaveToWord Files library, a workspace glossary lets you list names and terms so they are recognized more consistently in future transcripts.

Free transcription for specific situations

Different recordings call for slightly different approaches. Here is how we would handle the most common ones.

Interviews and podcasts

Record each person on their own microphone if possible, or at least position a single recorder evenly between speakers. Before transcribing, trim the pre-interview chat. After transcribing, check or add speaker names and tidy up false starts ("I, I think, um") only if you are producing a readable article; keep them if you need a verbatim record for research.

Lectures and classes

Ask permission before recording. Sit near the front or place your phone near the lecturer, away from the projector fan. Lectures often run longer than 15 minutes, so split the file into segments, transcribe each, and paste them together in order. Timestamps make it easy to jump back to a confusing explanation later.

Meetings

If the meeting software has its own recording, use that file rather than a phone recording of a laptop speaker. Remind participants that the meeting is being recorded. Afterward, if you want key points rather than every word, the Audio summarizer produces a summary with key points and action items alongside the full transcript.

Voice notes and dictation

Short memos are where free tools shine. Record in a quiet spot, speak in full sentences, and say punctuation-heavy items like lists slowly. If you want to record and transcribe in one place, the Online voice recorder records from your microphone in the browser and lets you send the recording on for transcription.

Audio from a video

If the speech you need is inside a video file, you do not have to extract the audio yourself. Upload the video to Video to text, which pulls out the soundtrack and transcribes it. Our guide on how to transcribe a video covers the details.

Privacy and security when you transcribe audio online

Recordings often contain personal information, confidential business details or other people's voices, so think about privacy before you upload.

• Know where processing happens. Online converters, including WaveToWord, process your file on remote servers rather than on your device. Built-in phone features vary: some process on-device, others send audio to the cloud for certain languages or features.
• Check retention. WaveToWord's tool files and results are stored privately, used only to produce your result, and deleted automatically after about an hour. Files you upload to your library stay until you delete them.
• Get consent. Recording laws differ by country and, in the United States, by state. Some require every participant to agree to being recorded. Tell people you are recording and why.
• Strip what you do not need. If only part of a recording is relevant, cut it down before uploading. Less data shared is less risk.
• Be careful with regulated data. Medical, legal and financial recordings may be covered by rules about how they can be processed. Check your organization's policy before using any online service.
• Secure your account. If you keep transcripts in an account, use a strong unique password and turn on two-factor authentication. WaveToWord supports optional authenticator-app 2FA in your security settings.

Honest limitations of free transcription

Free automatic transcription is excellent for drafts, notes, searchable archives and first-pass captions. It is not a replacement for a professional human transcriber when the stakes are high. Court records, medical documentation, published quotes and broadcast captions that must meet accessibility rules all deserve a careful human review, and sometimes a certified service.

Speaker labels are another weak spot. They are added only when the speech engine can tell voices apart, and even then two similar voices may be merged or one person split in two. Depending on the engine, a conversation may come back with every line under a single speaker. If who said what matters, check speaker changes yourself and rename or reassign them in the Files library editor. And no software can transcribe speech that was never captured clearly. If you can re-record, a better recording beats any amount of cleanup.

Quick checklist before you hit upload

Use this list every time and you will avoid most of the problems above:

1. Recording is the original file, not a re-shared copy.
2. Dead air and music at the start and end are trimmed.
3. File is under 15 minutes, or split into parts.
4. Format is a common audio type (MP3, M4A, WAV and similar).
5. Language is selected rather than left to chance, if you know it.
6. You have consent from the people in the recording.
7. You have set aside time to proofread names, numbers and quotes.

Start transcribing your audio for free

You do not need special software or a budget to turn speech into text. A reasonably clear recording, the right language setting and ten minutes of proofreading will get you a transcript you can trust for most everyday work.

Ready to try it? Sign in with your free account, open Audio to text, and upload your first recording. If your file is a video, use Video to text instead, and if you want the transcript of something already on YouTube, read our guide on how to get a transcript of a YouTube video.