How to make a movie in Arabic with AI: voice, dialects and subtitles

Video models paint silent pictures and can't reliably write Arabic letters inside a shot. So a generated Arabic film is built on two separate layers: shots with no speech and no writing, and on top of them an Arabic voice and correct Arabic subtitles. Getting that second layer right is the difference between a film that feels machine-translated and one your audience feels was made for them.

• A script written for the ear

  In MSA or a dialect, in short sentences that can be said in one breath.

• A voice in your audience's dialect

  MSA or one of 18 Arabic dialects with text to speech.

• Shots without writing

  Generated from text or an image in Vocla Video, with prompts written in English.

• Correct Arabic subtitles

  Joined letters, right to left, burned into the video or in an SRT file.

The dialect is a decision about your audience before it is a language decision. Modern Standard Arabic reaches every Arab viewer and suits historical, documentary and educational stories; a dialect makes viewers feel the film is talking to them. A short comedy for an audience in Riyadh lands better in Saudi Arabic, and for Cairo in Egyptian.

Vocla supports Modern Standard Arabic and 18 dialects: Saudi, Emirati, Kuwaiti, Qatari, Bahraini, Omani, Yemeni, Iraqi, Egyptian, Sudanese, Jordanian, Lebanese, Syrian, Palestinian, Moroccan Darija, Algerian Darija, Tunisian Darija and Libyan. Each dialect has its own page with sample scripts and writing tips.

Don't mix two dialects in the narrator's voice. If you want variety, give each character their own dialect and keep the narrator on one. Saudi and Emirati Arabic have detailed guides: Saudi Arabic voice generator and Emirati Arabic voice generator.

If you choose a dialect, write in it directly; don't write in MSA and expect it to turn into dialect. Spell words as they're said: «وش تبي؟» in Saudi, «عايز إيه؟» in Egyptian, «شو بدك؟» in Levantine. If your audience is mixed, write simple MSA close to everyday speech, not textbook Arabic.

Length: an Arabic narrator speaks about two words a second at a natural pace, and MSA is read a little more slowly than the dialects, so a 45-second film needs about 80 to 95 words. The free script planner estimates the length for MSA and for Saudi, Emirati, Kuwaiti, Egyptian, Jordanian and Lebanese Arabic.

• One sentence per idea; split long sentences in two.
• Write numbers and prices as words, the way they're said, such as «خمسة وعشرين ريال».
• Use punctuation to mark the pauses: commas and full stops help the voice breathe in the right places.
• Read every line out loud: if you stumble on it, so will the voice.

Dialogue turns a clip with a voice-over into a film. In text to dialogue you write the script one line per turn as "Name: line", pick a library voice for each character, and set a delivery for each line (natural, excited, calm, whisper, serious, happy, sad) and a pause after it, and you get one audio file with balanced voices.

A dialogue is spoken in one dialect: the dialect of its first Arabic voice. So if you want characters with genuinely different dialects, such as a grandfather speaking MSA, a grandson speaking Saudi and a friend speaking Egyptian, generate each character's lines in their dialect in text to speech or in a separate dialogue, then arrange them over the shots in your editor.

• Give each character a clearly different voice in pitch and age.
• Write each character's lines with their dialect's vocabulary, not in MSA.
• Generated shots don't sync lips, so use side-on or distant shots while characters speak, or a character on the phone.

Write scene prompts in English even when the film is Arabic, and add details of place and culture: an Arabic majlis with a coffee dallah, an old alley with lanterns, a busy street at night, dunes at sunset. In Vocla Video the English description for each scene is written automatically from your Arabic script, and you can edit it before rendering.

Don't ask for Arabic signs or titles inside the shot, because models garble Arabic letters; put the shop name or film title in the subtitles or the narrator's voice. If you want a character whose face stays the same, start their scenes from an image (Starter and up), and for product photos use the product video ad maker.

• Majlis: A traditional Arabic majlis with floor cushions and a brass dallah, warm evening lamp light, slow pan, cinematic
• Market street: A busy old market street at night with hanging lanterns and small shops, gentle handheld camera, warm and nostalgic
• Desert: Golden desert dunes at sunset with a camel caravan in the distance, aerial drone shot, epic and calm

Subtitles in an Arabic film serve people watching on mute, people who don't fully follow your dialect, and people who need to read the words. In Vocla Video the subtitles come from your own script, timed to the voice, in correct joined Arabic letters, in a Clean, Bold or Karaoke style.

For a finished film, upload it to AI captions: the speech is transcribed with timings, you review and edit the words, pick one of five styles, reframe to 9:16, 1:1 or 16:9, and export an MP4 with burned-in subtitles or SRT and VTT files. If the film is in English, the transcript can be translated into Arabic with each line keeping its timing.

In Vocla Video the subtitles match the script word for word, so a dialect voice gets dialect subtitles. If you want simple MSA subtitles over dialect audio so every Arab viewer can follow, edit the subtitle text in AI captions before exporting.

If you have a short film or a video in English, French, Spanish, Turkish, Hindi or Urdu, you can dub it into Arabic with AI dubbing: the speech is transcribed with timings, the speakers are detected, each speaker is cast with a matching voice, and the lines are adapted into the dialect rather than translated word for word.

Dubbing dialects: Saudi (Najdi or Hijazi), Emirati, Kuwaiti, Egyptian, Jordanian, Lebanese, Maghrebi Darija and Modern Standard Arabic. You review and edit every line in the studio, then export an MP4 with the Arabic track or with burned-in Arabic subtitles, and you can duplicate the project into another dialect in one click.

• Dubs are voice-over style, without lip-sync.
• The original voice isn't separated from the music; with Keep background, the original is lowered under the Arabic.
• Project length: dubbing is on paid plans only, up to 10 minutes per project on Starter, 30 minutes on Professional and 60 minutes on Business and Enterprise.
• Only dub what you own or have explicit permission to dub: publishing a dub of a commercial film without permission infringes copyright.

Choose instrumental music so it doesn't compete with the Arabic speech, and keep it clearly below the voice. In Vocla Video a music bed is added and ducked under the speech automatically, or you can generate a track from a text description with AI music.

Generate the film in the shape you'll post it in: 9:16 for TikTok, Reels and Shorts, 16:9 for YouTube and 1:1 for feed posts. In Vocla you get 9:16 from Starter and all three shapes from Professional, and the film comes as an MP4 with voice, subtitles and music, with any single scene regenerable on its own.

• A title in your audience's words

  Write the title and description the way people type their searches, in dialect if that's how your audience searches.

• A few precise tags

  Arabic tags that describe the story, the dialect and the genre, rather than dozens of generic ones.

• A clear cover

  A strong shot from the film with a short title added in the platform's app, in properly rendered Arabic.

• Safe zones

  Keep subtitles and faces away from the edges and bottom of the screen, where platform buttons appear. Details on the TikTok video maker page.

The whole path for a short film with an Arabic voice, a dialect and subtitles.

1. 1

   Choose the audience and the dialect

   Decide who the film is for, then choose MSA or that audience's dialect.

2. 2

   Write the script in the dialect and time it

   Write the lines the way they're spoken, and time them in the free script planner.

3. 3

   Add diacritics to ambiguous words and names

   Add tashkeel in one click and review the changes, and save hard names in the pronunciation dictionary.

4. 4

   Generate the film in Vocla Video

   Paste the script, then choose a library voice and the dialect, the shape, the Arabic caption style and music, and check the cost before you start.

5. 5

   Review the scenes and the storyboard

   Edit the English descriptions to fit an Arab setting, and approve a still for each scene before generating video.

6. 6

   Add character dialogue if there is any

   Generate the dialogue in text to dialogue with different voices, and place it over the right shots in your editor.

7. 7

   Export and post

   Download the MP4 and post it with a title and tags in your audience's words.

A short comedy of about 40 seconds called «المفتاح» (The Key), with a narrator and two characters. Each scene has its line in Egyptian Arabic and its MSA version, so you can compare how the feel changes between the two.

The voice demo on the text to speech page is free and needs no sign-up, up to 200 characters per try. The Free plan gives you ⁠500͏⁠ credits a month with no card for voice-overs, for personal use; dialogue, music, captions and dubbing come with every paid plan. Tashkeel and pronunciation tools are free on every plan.

Generating video shots, including starting from your own image, starts with Starter (videos up to 15 seconds, vertical, 720p, with a watermark), and all three shapes and 1080p without a watermark come with Professional. Commercial use, including monetised channels, is included from Starter up. Full details on the pricing page.

In some countries, currently including Egypt, audio generated on the Free plan ends with a short spoken tag saying it was made with Vocla. Paid plans never get the tag.

• Writing in MSA with a dialect voice

  The result is a hybrid that sounds odd. Write in the dialect you chose.

• Mixing dialects in the narrator's voice

  One narrator, one dialect; variety is for the characters.

• Leaving ambiguous words without diacritics

  One misread word is enough to weaken trust in the whole film.

• Arabic writing inside the shot

  The letters come out garbled. Titles and names belong in the subtitles.

• Subtitles that don't match the voice

  Decide whether subtitles are in the dialect or in MSA, and stick to it for the whole film.

• Dubbing content you don't own

  Only dub your own work or content you have explicit permission to dub.