How to transcribe an audio file to text: MP3, M4A, OGG and MP4
The file exists. It is on your phone, in the Downloads folder, in a meeting cloud, or it came over WhatsApp. What does not exist is the text: you need to find a passage, copy a sentence, send a clip to someone, or turn it into a document. And listening again is not an option.
Transcribing an audio file to text is the most direct path there is: you upload the file and get the speech written out, searchable by keyword, ready to edit. The problem is almost never the tool. It is a format that does not pass, a file that blows past the limit, or an upload that finishes and leaves you unclear where the text ended up.
This guide shows the full path: where to get the file, which formats are accepted, what the size limit is, and what to do with the transcript once it is ready.
Quick answer: how to transcribe an audio file to text
- Find the file. Phone recorder, downloads folder, WhatsApp media, or the meeting recording in the cloud.
- Check the format. MP3, M4A, WAV, AAC, OGG, OPUS and FLAC go straight in; videos in MP4, WEBM, MOV and MKV too.
- Open Sintesy and click New → Upload file. The panel also accepts drag and drop.
- Upload and watch the status. The file is processed and the transcript appears in the Transcript tab.
- Read the summary. Along with the text, Sintesy generates a structured summary of the content.
- Use the result. Search by keyword, copy passages, generate a mind map, checklist or questions from the same file.
First: find out where the file comes from
Half of upload problems happen before the upload: the file is somewhere you do not know how to reach.
| Where it comes from | Typical format | How to get the file |
|---|---|---|
| Phone recorder | M4A, AAC, MP3 | Open the recording app and use Share or Save to Files |
| WhatsApp (received audio) | OPUS inside OGG | Save the media to your phone or forward it to yourself and download |
| Meeting recording (Meet, Teams, Zoom) | MP4, WEBM | Download the recording from the platform cloud before ending the meeting |
| Video lesson or downloaded video | MP4, MKV, MOV | The file goes through the same flow: the transcription uses the audio track |
| Professional audio recording | WAV, FLAC, CAF | Export straight from the recorder or the editor |
WhatsApp audio arrives in OPUS, a format Sintesy accepts without conversion, so you do not need an intermediate app just to open the file. If that is exactly your case, the full flow is in transcribe WhatsApp audio and generate an automatic summary.
Formats Sintesy accepts
The import picker accepts audio, video and PDF, and the list of supported formats is long on purpose:
- Audio: MP3, WAV, M4A, AAC, OGG, OPUS, FLAC, AIFF, AMR, WMA, CAF, 3GP
- Video: MP4, WEBM, MOV, MKV, AVI, MPEG, M4V, FLV, WMV
- Document: PDF
If the file is not recognized, the panel answers with the exact message: “Media format not supported”.
Three behaviors worth knowing:
- Video goes in through its audio. You do not need to separate the track: the video file is uploaded and the text comes from the speech.
- PDF also goes through the same button. For scanned pages, the system uses OCR when the page text is not selectable.
- Plain text also becomes a summary. When you paste long text into the panel, Sintesy asks “Create a sintesy from this text?” before processing.
Limits that apply before you send
| Limit | Free plan | Pro plan |
|---|---|---|
| Size per audio or video file | 20 MB | 2 GB |
| Size per PDF | 64 MB | 64 MB |
| Duration processed per day | 60 minutes | No daily limit |
| Summaries per day | 10 | No limit |
The size limit shows up as an error on the upload itself: “Upgrade your plan to upload files larger than 20MB!”. The daily limits reset every day, so a large audio that did not fit yesterday may fit today, if you still have minutes available. The Pro plan removes both limits.
In practice, 20 MB covers tens of minutes of MP3, but it depends on the format: a file at 128 kbps yields about 20 minutes, while a mono WAV of just a few minutes already exceeds that.
The flow in the dashboard: New → Upload file
- Open the dashboard and click New → Upload file.
- Drag the file to the area that says “Drag and drop audio, video, or PDF files here”, or pick it with the file picker. You can also drop the file anywhere on the panel.
- Track the processing status. The file goes through transcription and then content generation.
- Open the Transcript tab to read the text utterance by utterance.
- Read the summary, which condenses the content into structured blocks.
- Generate the derived formats when it makes sense: a mind map to see the structure, a checklist to turn agreements into tasks.
The text and the summary are saved in your account, searchable from the panel. You do not need to keep the file open or store the original to consult the content later.
When the audio is too long
Long audio is where most people get stuck. Three paths:
- Split the recording into parts. Cut the file into blocks of up to 60 minutes and send one per day on the free plan. Each part becomes its own summary.
- Choose a compressed format. Converting a WAV to MP3 reduces the size without changing the content and avoids blowing past the 20 MB.
- Focus on the passage that matters. If you only need the final decision or a fifteen-minute block, send just that passage. Less file, less processing, faster response.
If the goal is turning the whole audio into study material or team documentation, the Pro plan removes the size limit and the daily limit.
What to do after the text
The transcript is the raw material, not the end of the flow. From the same file:
- Summary: Sintesy already delivers the condensed reading. For long text pasted into the panel, the path is summarize long text with AI.
- Mind map: shows the structure of what was said, useful for class, talk and long meeting.
- Checklist: turns agreements into tasks with an owner, like in the meeting checklist.
- Search: finds the decision, name or deadline without re-listening to anything, as in searchable transcripts for finding decisions.
- Spot-check: go back to the audio only in the passages where the text looks wrong.
That last point matters: AI transcription gets names, numbers and technical terms wrong when the audio is bad. If the result looks strange, the first suspect is the recording, not the model, as explained in wrong transcript: the problem is almost always the audio.
Five mistakes that block the first upload
- A format that is not on the list. ARF, a proprietary recorder or an exotic extension: convert to MP3 before sending.
- File above the limit. 20 MB is the free plan ceiling for audio and video; for PDF, 64 MB.
- Sent the wrong folder. On a phone, check that what you selected is the audio file and not a shortcut or a .nomedia file.
- Audio that is only noise or silence. Transcription does not invent speech that is not there. If the file has three hours of silence, the text comes out empty.
- Waiting for the text in the wrong place. After processing, the content lands in the Transcript tab of the created item, not in an email or a notification.
If you have not used the full flow yet, the complete guide to audio transcription with AI shows the method in five steps, from file to application.
Frequently asked questions
What is the best format for transcribing audio?
Any accepted format works, but MP3 and M4A offer the best balance of size and quality. If the original is WAV or FLAC, the file gets bigger with no real gain for the transcription: convert to MP3 before uploading.
Do I need to install anything?
No. The whole flow happens in the browser: you upload the file through the panel and get the transcript in the same account.
Can I transcribe WhatsApp audio directly?
Yes. The audio arrives in OPUS, a format the importer accepts. Save the media to your device and upload it, as described in transcribe WhatsApp audio with a summary.
What if the video file is large?
Video follows the same rule as audio: 20 MB on the free plan and 2 GB on Pro. For classes and meetings recorded in MP4, download the lower-resolution version or split it into parts.
Does the transcript replace listening again?
To find information, yes: keyword search resolves it. To judge tone, hesitation, or what was said between the lines, go back to the audio. The text is the entry point, not the final proof.
The short path
Find the file, check the format, upload via New → Upload file, and read the summary before the full text. Use the transcript to search and copy, the mind map to understand the structure, and the checklist to turn an agreement into a task.
The gain is not “never listening to the audio again.” It is reaching the right passage in seconds and turning a dormant file into something you consult, edit and share.
