Product Video from Photo with AI: Turn One Product Shot into an Arabic Video Ad

01How AI makes a product video from a photo

You have a good shot of your product, a perfume bottle, a coffee bag, a handbag, and you want a short ad for Reels or TikTok. The old route was a shoot, an edit and a studio voiceover. In Vocla, the photo itself becomes the starting point: you upload it as the first frame, the video model animates it and builds the scene around it, and the voice and captions come from tools built for Arabic.

Understand the division of labour up front. The video model only paints moving pictures; it doesn't speak and it doesn't write. Every word the viewer hears is a Vocla voice in the dialect you chose, and every word they read is a separate caption track. That split is why the Arabic comes out right: video models garble Arabic letters when asked to draw them in frame.

02Prepare the photo before you upload

The first frame sets the quality ceiling for everything after it. A dark, cluttered photo gives you a dark, cluttered video. Before uploading, check:

• Even light: soft daylight from the side, no hard shadows hiding product detail.
• Simple background: wood, marble or a single colour, so the model can tell product from setting.
• Product centred with breathing room: camera moves need margin around the subject.
• Shape matches the ad: for a vertical Reel, a portrait or square photo beats a wide landscape crop.
• Little text in frame: small labels and packaging copy can warp as the scene moves, so don't rely on them to carry the message.
• Sharp source: a crisp phone photo is plenty, as long as it isn't a compressed screenshot.

03Write the ad script first, in dialect

Vocla Video starts from the script, not the picture. It voices the script first, then splits it into scenes of 3 to 8 seconds timed to the speech. So write three beats: a hook, one clear benefit, a call to action. A Saudi script for specialty coffee, about 15 seconds:

An Egyptian perfume ad with a lighter touch:

And an English version for bilingual campaigns, written as an English ad rather than a translation:

04Prompt each scene: direct the motion, not the product

After voicing, Vocla writes an English visual prompt for each scene, which you can edit. With your photo as the first frame, the prompt should describe motion, light and camera rather than the product, because the product is already in the frame. Patterns that tend to work:

• One move per scene: an orbit plus a zoom plus a lighting change confuses the model.
• Never ask for text: leave store names and prices out of prompts; they belong in captions.
• No real people: if you need a hand holding the product, ask for a generic person.
• Redo one scene: if a scene misses, regenerate that scene alone instead of the whole video.

05A worked example and common mistakes

Take the Saudi coffee script above. After voicing, Vocla typically produces a scene plan like this one, and every row can be edited before you spend credits on video:

The last scene is deliberately calm, with empty space at the bottom of the frame, because that is where the captions and call to action sit. None of the prompts mention the store name or price; that is the job of the voice and the caption track.

These are the mistakes we see most often from first-time users, and all of them are easy to avoid:

• Long script, short ad: fifty words in fifteen seconds makes the voice rushed and the scenes choppy.
• Busy background: shelves and clutter around the product move with it and pull the eye away.
• Skipping the storyboard: rendering straight away means discovering weak scenes after paying for them.
• Mixed dialects: a Saudi voice over MSA captions feels stitched together; keep captions matching the speech.
• Relying on the logo in frame: say the brand name and show it in captions, and let the picture carry mood and product.

06The full workflow in the Video studio

1. 1Script and voice: paste the script, pick voice and dialect, then format, quality tier, caption style and music.
2. 2Product photo: on Starter and up, upload it as the first frame.
3. 3Live quote: see the credits and the seconds left this month before anything renders.
4. 4Scene plan: check each scene's timing and spoken line, and edit its prompt.
5. 5Storyboard: generate a still per scene and approve the look before paying for video.
6. 6Render and download: scenes render, voice, captions and music are mixed, and you download the MP4 or redo a single scene.

Arabic captions are a separate track timed to the speech, shaped correctly and set right to left, in clean, bold or karaoke style at the bottom or centre of the frame. That matters because most Reels are watched muted.

07Cost and which plan fits

Video is billed per rendered second by quality tier, plus 200 credits per generated storyboard still (scenes that use your uploaded photo aren't charged for one), 120 credits for a music bed, and the voiceover. A 15-second Standard ad at 400 credits per second is about 6,000 credits for the scenes, and failed scenes are refunded.

For a store publishing product ads every week, Professional at $29 a month is the practical choice: 80,000 credits, 300 video seconds a month, 1080p and no watermark. Starter at $4.99 is for testing the idea: an ad from your product photo up to 15 seconds, vertical 720p with a watermark.

Product names in the scripts are fictional. Credit rates are current defaults and can change; the exact price always appears in the studio quote before you render.