Most caption tools are built for English first. Persian support is added later, and it shows. If you have ever exported a Reel and found your Persian captions scrambled, broken apart or set in a font that looks nothing like the preview, you are not alone.
This post explains where things usually go wrong, gives you a checklist you can use with any tool, and shows how capshak deals with each problem.
Why Persian captions break
Persian is written right to left, its letters join, and it shares space with English words and numbers all the time. Every one of those facts trips up software that assumed left-to-right Latin text. The usual symptoms:
- Letters that don’t join. The renderer draws each letter in its isolated form, so words look like a row of separate shapes.
- Punctuation on the wrong side. A period or question mark jumps to the start of the line.
- Mixed text out of order. An English brand name or a number in the middle of a Persian sentence pushes the words around.
- Missing half-spaces. You get «می شود» or «ویدیو ها» instead of the correct joined-but-separate forms.
- Font fallback. The chosen font has no Persian glyphs, so the system swaps in a default that ignores your style.
- Bad line breaks. A short phrase is split across two lines, or a line ends on a lonely preposition.
Right to left is more than flipping text
A right-to-left caption is not an English caption mirrored. The whole block has a direction: where lines start, which side punctuation sits on, how a word-by-word highlight moves across the line. When a tool only reverses characters, the result may look fine for one plain sentence and fall apart as soon as there is a comma, a number or a Latin word.
Watch the highlight in a karaoke-style caption. In a correct Persian caption it moves from right to left, the same way the viewer reads.
The half-space problem
The zero-width non-joiner (ZWNJ), called nim-fasele in Persian, keeps two parts of a word visually close without joining their letters. It belongs in verb prefixes like «میشود», plural endings like «ویدیوها» and «زیرنویسها», and many compound words.
Speech-to-text models often get this wrong, because the sound is the same with or without it. A caption full of plain spaces where half-spaces should be looks careless to a native reader, even if every word is technically correct.
Mixed Persian, English and numbers
Real speech mixes languages. Creators say product names, app names and English terms in the middle of Persian sentences. A few rules help:
- Keep brand names in their original spelling if your audience knows them that way.
- Pick one style of digits for a video and stick to it. Persian digits read naturally in Persian captions.
- Keep a mixed phrase on one line when you can. Breaking a line right at the switch between scripts is where order problems show up most.
Fonts that actually support Persian
A font that “supports Arabic” is not always good for Persian. Check that it has the four Persian-specific letters (پ چ ژ گ), Persian digits and the Persian forms of ک and ی. Then check how it looks at caption size on a phone: thin, decorative fonts can be beautiful in a poster and unreadable over moving video.
Display fonts like Lalezar or Titr work for short, punchy captions. Clean text fonts like Vazirmatn or Estedad are safer for longer lines.
Line breaking and length
Short lines read faster. One or two lines on screen at a time is a good default for short-form video. Break at natural pauses in the sentence, not in the middle of a phrase, and keep the subject and its verb together when you can.
A practical checklist
Before you export Persian captions, check:
- Letters join correctly in every word.
- Punctuation sits at the end of the sentence, on the left.
- Half-spaces are in place: «می», «ها», «ای», compound words.
- Names and terms are spelled the way you want them.
- Mixed English words and numbers appear in the right order.
- The font has full Persian coverage and is readable at phone size.
- No more than two lines on screen, broken at natural pauses.
- The preview matches the exported file.
How capshak handles Persian captions
capshak was built with Persian as a first-class language, not an add-on.
- Word-timed captions. Every word has its own timing, so word-by-word highlights stay in sync with the speaker.
- AI proofreading. One click fixes misheard words and names, spelling, half-spaces and punctuation, while keeping the words that were actually spoken. You can add a hint with names or terms so they come out right. Proofreading costs 1 credit per 2 minutes of video.
- Persian fonts included. 48 bundled Persian fonts, including Vazirmatn, Estedad, Lalezar, XM Yekan, Titr and Iran Nastaliq.
- Styles made for Persian. Of the 24 ready-made caption styles, more than 8 are designed for Persian text.
- Your own fonts. With “My fonts” you can add a font you own (TTF, OTF, WOFF or WOFF2). It stays on your device.
- Same picture in preview and export. The preview and the export use the same renderer, so what you check is what you get. You can also export an SRT file.
The interface itself is fully bilingual, so you can work in Persian from start to finish.
Try it on your next video
Upload a short clip, generate captions, run proofreading and pick a Persian style. Editing and exporting are free, and captions come with any plan. Create your free account, or browse the caption styles first.