How to Add Arabic Captions to Reels and TikTok That Follow the Speech and Read Right
How to add Arabic captions to Reels and TikTok that follow the speech: burned-in vs platform captions, styles, safe zones and fixing broken Arabic text.
In this article
To add Arabic captions to Reels, transcribe the audio with word-level timings, correct the text, pick a clear style and burn the captions into a 9:16 video, keeping them above the platform's buttons. In Vocla's AI captions you do all of this in one place, with Arabic rendered right to left and properly joined, and five styles including karaoke.
01Why captions matter on Reels and TikTok
A lot of people scroll Reels and TikTok with the sound off: at work, in a waiting room, or late at night next to someone asleep. Without captions, your message is lost however strong the content. Captions also help viewers who are hard of hearing, viewers following in a second language, and anyone checking a product name or a price.
Arabic captions bring their own problems. Many tools treat Arabic like English: letters come out disconnected, words appear in reverse order, or the type is much smaller than the Latin next to it. This guide covers how to avoid all of that, and focuses on the captions themselves. Writing the voiceover for your Reels has its own guide.
02Burned-in captions or platform captions?
There are two ways to add captions, and each has its place:
| Point | Burned into the video | Platform captions or SRT file |
|---|---|---|
| Visibility | Always shown, on every device and platform | Viewer can switch them on or off |
| Look | Exactly as you designed it | Platform's own styling |
| Editing after posting | Needs a new render and re-upload | Edit the file or text |
| Best for | Reels, TikTok, short-form | Long YouTube videos, players |
| Several languages | One video per language | One file per language on the same video |
For Reels and TikTok we recommend burned-in captions, because they show exactly as designed even when viewers never tap a captions button. Where a platform accepts subtitle files, upload an SRT as well. For long YouTube videos a subtitle file is usually better, because viewers control it.
Burned-in captions are part of the picture and can't be switched off. Don't burn in a typo hoping to fix it later; correct the transcript before you render.
03Caption styles: which one fits your content?
Style changes the feel of the whole video. Vocla's AI captions offer five, each with a job:
| Style | Look | Best for |
|---|---|---|
| Clean | White text with an outline | Most videos; readable on any background |
| Boxed | Each line on a translucent box | Busy or bright footage |
| Bold | Larger, heavier type, thick outline | Short-form and strong hooks |
| Karaoke | Boxed, each word lights up as spoken | Reels, TikTok, high-energy content |
| Minimal | Smaller text, soft shadow, no outline | Calm, polished content |
Karaoke is the look of the moment on Reels, because the highlighted word pulls the eye and keeps people watching to the end. It follows the recognised word timings, so it works on captions in the video's own language. For translated captions, such as Arabic on an English video, use one of the other four styles.
- Position: bottom, middle or top.
- Size: small, medium or large.
- Colours: text colour and highlight colour for the current word.
- Lines: a maximum per caption; one or two lines suit Reels best.
- Speaker names: on or off, useful for interviews and podcast clips.
04Safe zones: where captions should sit
The mistake we see most: good captions buried under the account name and description at the bottom, or covered by the like and comment buttons on the side. Reels and TikTok overlay their interface on your video, so captions need to stay clear of it.
- Lift them off the bottom: the lower part of the screen carries the handle, caption text and audio label, so sit captions clearly above it.
- Keep off the edge: the action buttons run down one side, so keep lines centred and away from the edges.
- Don't hug the top: some platform elements appear there too.
- Check on your phone: post as a draft or private and view it in the app before going public.
In Vocla, when you reframe to 9:16, captions on vertical videos are lifted above the platform's buttons automatically. A 16:9 source becomes 9:16 with a centred crop, or Vocla keeps the whole picture over a blurred fill when a crop would lose too much. Reframing doesn't follow faces or moving subjects.
Platforms update their interfaces regularly, so we won't give fixed pixel figures. The reliable rule: preview inside the app itself and make sure no word sits under a button or text.
05Arabic text problems and how to avoid them
If you've typed Arabic into some editors or AI tools, you've probably seen at least one of these:
- Disconnected letters: each letter stands alone instead of joining, because the tool doesn't shape Arabic.
- Reversed words: the line is laid out left to right, so words appear in reverse order.
- Mixed-script chaos: an English word or a number inside an Arabic line scrambles the order.
- Tiny type: Arabic renders visibly smaller than Latin at the same size.
- Garbled text in AI scenes: AI video models mangle Arabic if you ask them to draw it inside the frame.
خصم 20% على أول طلب لك، بس اليوم!20% off your first order, today only!
A line like this, with a number, a percent sign and an exclamation mark, is exactly what trips up tools without proper Arabic support. In Vocla, captions render right to left with correctly joined letters, words never reverse, and the font is sized to match English so mixed lines look even. In Vocla Video, captions are a separate track burned over the scenes; we never ask the video model to write them.
06Step by step in Vocla
- 1Upload: in AI captions, upload the video, or audio only, which becomes a captioned video with a waveform background.
- 2Check the words: Vocla transcribes with speakers and word timings. Fix names and spellings in the editor before rendering.
- 3Pick a style and frame: style, position, size, colours and line count, then reframe to 9:16. The live preview shows the result.
- 4Render and download: the MP4 with burned-in captions lands in your Library; download SRT or VTT for platforms that take subtitle files.
- 5Another language: burn in the translation instead of the original, such as Arabic captions on an English video, or English on an Arabic one.
Editing lines or options after a render marks it out of date; render again to update the MP4.
Read alsoAI captions for Arabic video07Writing captions people can read at a glance
Viewers get a second or two per caption before it changes, so how you break lines matters more than the font. The rules we apply:
- Short lines: three to six words per line, two lines maximum on screen.
- Break at the pause: split where a thought ends or the speaker breathes, never inside a name or a number.
- Consistent dialect spelling: if you write a word one way, keep it that way for the whole video.
- Digits on screen: “20%” reads faster than the spelled-out form, even if it's spoken in words.
- Correct names: fix product and people's names in the editor; they're what viewers notice first.
- Contrast: on bright or busy footage, switch from Clean to Boxed.
Each line should be one short idea you can read in a single look, and a question can break where it builds curiosity for the next line. If you also need an Arabic SRT file for another platform, preparing it and checking its text direction has its own guide.
08What Arabic captions cost
Captions start from a transcript, so the video is transcribed first and then rendered. Per Vocla's AI captions page, transcription is 150 credits per started minute, the caption render 30 credits per started minute, and translated captions add 60 credits per minute. Both costs are shown before you start.
| Video | Transcription | Render | Total |
|---|---|---|---|
| 45-second Reel | 150 | 30 | 180 credits |
| 90-second Reel | 300 | 60 | 360 credits |
| 5-minute video | 750 | 150 | 900 credits |
| 5-minute video, translated captions | 750 + 300 | 150 | 1,200 credits |
So AI captions are on every paid plan, with all five styles and reframing, but not on the Free plan. A 45-second Reel uses 180 credits, so Starter's 5,000 monthly credits cover about 27. The longest file is 15 minutes on Basic, 30 on Starter, 120 on Professional and 300 on Enterprise.
Key takeaways
- Burn captions into Reels and TikTok, and add an SRT file where the platform accepts one.
- Karaoke grabs attention but works on original-language captions; use the other styles for translations.
- Keep captions above the handle and description and away from the side buttons; preview in the app.
- Vocla renders Arabic right to left with joined letters, and a 45-second Reel costs about 180 credits.
Frequently asked questions
How do I make captions follow the speech on Reels?
You need a transcript with a timing for every word, then captions displayed on those timings. Vocla's transcription gives you word timings, and the karaoke style lights up each word as it's spoken.
Why do my Arabic captions look broken or reversed?
Some tools don't join Arabic letters or don't lay text out right to left. Use a tool with real Arabic support, and don't ask AI video models to draw text inside the scene.
Should I burn in captions or upload an SRT file?
For Reels and TikTok, burn them in so they always show as designed. SRT suits long YouTube videos because viewers control it, and you can do both.
Can I add Arabic captions to an English video?
Yes. Vocla translates the transcript and you can burn in the Arabic translation instead of the original. Translated captions add 60 credits per minute and use every style except karaoke.
How much do captions cost for a one-minute Reel?
150 credits for transcription and 30 for the render, so 180 credits per started minute, per Vocla's AI captions page. The cost is shown before you start.
- #Arabic subtitles
- #burned-in captions
- #karaoke captions
- #SRT files
Read next
- Speech to text & captionsOctober 3, 2026 · 7 min read
How to Make an Arabic SRT File for Your Video: From Transcript to YouTube Upload
Read article - Speech to text & captionsOctober 3, 2026 · 6 min read
Arabic Dialect Speech Recognition: Why It's Harder Than MSA and How to Improve It
Read article - Speech to text & captionsOctober 3, 2026 · 7 min read
Arabic Meeting Transcription: From a Recording to Written Minutes, Decisions and Action Items
Read article