Speech to textAvailable now

Turn Arabic audio into text you can use

Upload a recording or a video. Vocla detects the language and who speaks when, writes every word with its time and can translate the transcript. Then you edit, search and export it in the format you need.

Transcribe your first file See the exports

Speakers detected automatically
Word-level timestamps
Word, SRT, VTT, TXT, JSON and CSV

export formats, from Word to CSV: 6
credits per started minute of audio: 40
per file on Professional: 120 min
language and speaker detection: Auto

How it works

From recording to transcript in four steps

Most files are ready in a few minutes, and you see the cost before you start.

Step 1
Upload audio or video
Drop in an interview, a lecture, a meeting or a video, up to 500 MB per file.
Step 2
Pick the language, or let Vocla detect it
Arabic, English and French are recognised directly, other languages through AI detection. Add names and terms so they are spelled your way.
Step 3
Review in the editor
Playback follows the text word by word. Fix a word, rename a speaker, split or merge lines, or find and replace across the file.
Step 4
Translate and export
Translate the transcript side by side, then download Word, SRT, VTT, TXT, JSON or CSV with the original, the translation or both.

Accurate by design

Built for real Arabic recordings

Interviews, lectures and calls rarely come with one clear voice. Vocla keeps track of who said what and when.

Who spoke when
Speakers are detected automatically and every line is labelled, so an interview reads like a script. Rename Speaker 1 to a real name once and it changes everywhere.
Every word has a time
Word-level timestamps drive the editor's highlighting, precise subtitle timing and karaoke captions.
Arabic and its dialects
Pick Modern Standard Arabic or a dialect, or leave it on auto: Arabic, English and French are detected directly, other languages through AI detection.
Names spelled your way
Add a list of names and terms (people, places, brands) to improve how they are recognised. In the demo, the podcast's name and both hosts' names came out as listed.

The editor

Edit it like a document, check it like a recording

Everything happens against the audio, so checking a line takes a second.

Synced playback
The line being spoken and the word inside it are highlighted as the audio plays.
Click to seek
Click any line or word to hear it from there.
Inline editing
Correct a word where it is. Timings stay attached to the line.
Speaker rename
Turn Speaker 1 and Speaker 2 into real names for the whole transcript.
Split and merge
Cut a long line where a new sentence starts, or join two short ones.
Find and replace
Fix a spelling or a name across the whole file in one go.

Translation

Translate the transcript, side by side

Translate English to Arabic, Arabic to English and more. Each line keeps its time, so the translation exports to subtitles as easily as the original, or next to it in a bilingual file.

Translation adds 20 credits per minute.

Exports

Take the transcript anywhere

Six formats, each with the original, the translation or both. The excerpts below are the real files exported from the demo.

Numbered cues with start and end times, wrapped for the screen, with Arabic marked right to left.

First lines of the exported file

1
00:00:00,400 --> 00:00:02,700
Speaker 1: Welcome back to the show today.

2
00:00:02,700 --> 00:00:05,500
Speaker 1: We're talking about
starting a small business.

Original, translation or both
Speaker names on or off
Timestamps on or off

Use cases

Who uses speech to text

Anyone who records people talking and needs the words in writing.

Journalists
Transcribe interviews and press conferences, find the quote by searching, and check it against the audio in a click.
Podcasters
Show notes, episode transcripts and subtitle files for clips, with each host named.
Researchers
Turn focus groups and field interviews into searchable text, and export CSV or JSON for analysis.
Call centres
Transcribe recorded calls with agent and customer on separate lines, for quality reviews and training.
Creators
Subtitles and captions for every video, in Arabic and in translation.
E-learning
Lecture transcripts and subtitles that make courses searchable and accessible.

For teams

Transcripts that fit how your team works

Speech to text runs inside your Vocla workspace, with the same roles, approvals and budgets as the rest of the product.

RolesViewers can open, search and export transcripts; editors upload, edit, translate and render.
ApprovalsWhen your workspace requires approval, caption renders wait for an approved review before anyone downloads them.
Credit capsMember credit caps cover transcription, translation and caption renders.
Client tagsTag a transcript with a client and its charges and caption renders follow, for clean usage reports.

Plans and credits

Clear prices, clear limits

Transcription is charged per started minute, from the file's length, and the cost is shown before you start. If a transcript fails, its credits are refunded.

Clear prices, clear limits
Plans and credits	Longest file	Translation	All six exports
Free	10 minutes
Professional	120 minutes
Enterprise	300 minutes

Credits

Transcription: 40 credits per started minute
Translation: +20 credits per minute
Upload size: Up to 500 MB per file
Example: a 30-minute interview: 1,200 credits (1,800 with translation)

Compare plans

Guide

How to transcribe Arabic audio with AI

How Arabic speech to text works

Speech to text listens to a recording and writes down what is said, with the time of every word. Vocla first finds the language (or uses the one you choose), then recognises the speech, splits it into lines at pauses and sentence ends, and works out which speaker says each line.

The result is not a wall of text: it is a transcript you can play, search and edit, with each line tied to its moment in the audio. That is what makes Arabic transcription useful for subtitles, quotes and research, not only for reading.

Getting the most accurate Arabic transcription

Clear audio matters more than anything else. Record close to the speaker, avoid music under speech, and give each person their own turn. If you know the dialect, choose it; if the clip is short or mixes languages, choosing the language yourself is more reliable than auto-detect.

Names are where any recogniser struggles, so add the people, places and brands in the recording to the names list before you start. After that, a pass in the editor with find and replace usually takes a few minutes, even on long files.

Transcripts, subtitles or captions?

A transcript is the full text with speakers and times, for reading, quoting and searching. Subtitles are the same words cut into short timed cues (SRT or VTT) for a player. Captions are subtitles burned into the picture, so they show everywhere, including on social feeds with the sound off.

In Vocla they all come from one transcript: fix a word once and the Word file, the subtitles and the captioned video all use the corrected line.

Translating a transcript

Translation runs line by line with the surrounding lines as context, and keeps each line's time. You can read it next to the original, export it alone, or export both in one bilingual file, which is handy for reviewers who read only one of the two languages.

Translate English interviews into Arabic for an Arabic audience, or Arabic recordings into English for partners and research. Translation adds 20 credits per minute to the 40 of transcription.

What a transcript costs

Transcription uses 40 credits per started minute, taken from the file's length before it starts and refunded if it fails. A 30-minute interview uses 1,200 credits, or 1,800 with a translation. The Free plan's 10,000 monthly credits cover about four hours of transcription, with files up to 10 minutes each.

Questions about Speech to text

How accurate is Arabic speech to text?

On clear recordings the words are usually right, with Arabic spelling (hamzas, ة and ى) restored; the demo on this page is published unedited so you can judge for yourself. Expect to check names and add some punctuation, since transcripts come without tashkeel and punctuation inside a line can be sparse. The editor makes that a quick pass.

Which languages and dialects can I transcribe?

Arabic, including dialects you can choose, plus English, French and other languages. On auto, Arabic, English and French are detected directly and other languages through AI detection.

Does it detect different speakers?

Yes. Speakers are detected automatically and every line is labelled Speaker 1, Speaker 2 and so on. You can rename speakers and move lines between them in the editor.

Can I transcribe a video, not just audio?

Yes. Upload audio or video; Vocla transcribes the soundtrack. The same transcript can then become burned-in captions with AI captions.

How long can a file be?

Up to 10 minutes per file on Free, 120 minutes on Professional and 300 minutes on Enterprise, and up to 500 MB per upload.

How much does transcription cost?

40 credits per started minute of audio, plus 20 credits per minute if you translate. The cost is shown before you start and refunded if the transcript fails.

Can I translate the transcript?

Yes, for example English to Arabic or Arabic to English. The translation keeps each line's time and appears side by side with the original, and you can export the original, the translation or both.

Which export formats are there?

Word (DOCX) with proper right-to-left Arabic, SRT and VTT subtitles, plain text, JSON with word timings, and CSV that opens correctly in Excel.

Do Arabic transcripts open correctly in Word and Excel?

Yes. The Word file uses right-to-left paragraphs for Arabic lines, and the CSV is saved as UTF-8 with a byte-order mark so Excel shows Arabic properly.

How do I improve how names are recognised?

Add them to the names and terms list when you start the transcript. Vocla uses the list to recognise them and writes a matching word the list's way.

Is there live, real-time transcription?

No. Speech to text works on uploaded files, usually ready within a few minutes.

Can my team work on transcripts together?

Yes. Transcripts live in your workspace: viewers can read and export, editors can edit and translate, and member credit caps and client tags apply.

Turn Arabic audio into text you can use

From recording to transcript in four steps

Upload audio or video

Pick the language, or let Vocla detect it

Review in the editor

Translate and export

Built for real Arabic recordings

Who spoke when

Every word has a time

Arabic and its dialects

Names spelled your way

Edit it like a document, check it like a recording

Synced playback

Click to seek

Inline editing

Speaker rename

Split and merge

Find and replace