If you need the original-language wording, choose Transcribe under Task. If you need an English version, choose Translate to English.
Overview
Need notes from a meeting, interview or lecture? Upload the recording, turn the speech into text, and check the wording before you share it.
How to use AI Audio Transcriber
1. Upload your recording
In Source, choose Upload audio/video to transcribe. Select one file up to 100 MB. Accepted formats include MP3, WAV, OGG, FLAC, M4A, AAC, WMA, MP4, WebM and MKV.
2. Choose the language and task
Set Speech Language if you know it, or leave Auto-Detect selected. Choose Transcribe or Translate to English under Task.
3. Choose an output format
Select Plain Text for a readable transcript, SRT Subtitles or WebVTT for a subtitle workflow, or JSON for structured output. Include timestamps and Word-level timestamps are separate controls; choose the timing detail your task needs.
4. Process, review and save
Click Process. Read the result while listening to the relevant parts of the recording. Check names, numbers and unclear passages before using Copy or Download. If you chose subtitles, open the downloaded file with your video and check its timing.
Choose the setting for your task
| What you need | Where to start |
|---|---|
| Meeting notes to read or edit | Choose Plain Text. |
| Subtitles for a video | Choose SRT Subtitles or WebVTT to match your player or editor. |
| Structured data for another workflow | Choose JSON and inspect the returned fields. |
Limits to know
Verify important wording
Use the recording to resolve uncertain words. Do not turn an unclear name, amount or instruction into a confident quotation without checking it.
Translation changes the task
Translate to English produces an English version. If you need the original-language wording, select Transcribe instead.
Settings reference
| Setting | Options | Default |
|---|---|---|
| Speech Language | Auto-Detect, English, Spanish, French, German, Italian, Portuguese, Japanese, Korean, Chinese, Arabic, Hindi, Russian | auto |
| Include timestamps | Yes / No | True |
| Task | Transcribe, Translate to English | transcribe |
| Output Format | Plain Text, SRT Subtitles, WebVTT, JSON | text |
| Accuracy | 1–5 | 1 |
| Word-level timestamps | Yes / No | False |
Examples & use cases
Someone preparing a transcript or subtitles
Prepare interview notes
You need a draft transcript to check against a recording.
One supported recording up to 100 MB to Plain text
- Choose the known Speech Language and Task: Transcribe.
- Select Plain Text and process the recording.
- Replay names and quotations, correct the text, then copy or download it.
Expected outcome
The transcript is a draft; verify speaker attribution yourself.
Someone preparing a transcript or subtitles
Prepare subtitle timing for a lesson
Your editor needs a subtitle file for recorded speech.
MP4 with spoken audio to SRT subtitles
- Set Speech Language and choose SRT Subtitles.
- Process and download the file.
- Load the SRT beside the original video and check wording and timing.
Expected outcome
A subtitle file needs review in its destination player or editor.
Access & cost
Sign-in required. Transcription uses credits, with cost affected by file size. Review the displayed cost before processing.
Privacy
Check the processing mode before starting. Online processing uploads files; completed results may remain available for later download. See the Privacy Policy for current handling and retention details.
Common questions
Can I use a video recording?
Yes. MP4, WebM and MKV are accepted alongside the listed audio formats.
Will a transcript identify who said each line?
Do not rely on speaker identification for this workflow. Confirm speaker attribution against the recording before adding it to your notes.
What should I check before sharing an automatic transcript?
Replay names, numbers, unfamiliar terms, and sections with overlapping speakers. Automatic text can sound fluent while mishearing a word. Correct the transcript against the recording before using it as a quotation or a record of what someone said.