How to make a product video with AI — no filming

What you need before you start

One product page or 1–3 sharp photos, plus three facts worth saying out loud. Every shot and every line of the video comes from these, so they matter more than any editing trick.

• Photos: the product in focus, well lit and filling a good part of the frame, on its own, with no other products or people. Different angles, such as the front, a side and a close-up detail, give you shots to cut between.
• Clean images: no text overlays, badges or collages. They end up in every shot, along with the product.
• Size: at least 1080 pixels wide, the width of a 1080×1920 video. Smaller photos get enlarged and look soft.
• Three facts from your product page: what it does, one number (size, capacity, how long it lasts) and how to use it. The script needs nothing else.
• Rights: photos, music and words that are yours or licensed for ads. Use a supplier's photos only if you're allowed to use them in ads.

To do it by hand you also need a way to animate photos (an AI image-to-video tool, or the zoom and pan in any video editor), a voice (your own recording or text-to-speech), music licensed for ads, and an editor that burns in captions and exports MP4.

Pick a hook for the first 3 seconds

Say what the product is and why it matters before the 3-second mark, with the product on screen from the first frame. TikTok's creative guidance for ads asks for the content proposition in the first 3 seconds and the hook within the first 6 (TikTok creative best practices, updated June 2025).

A hook has two parts that carry one idea: a line of text on screen, up to about eight words, and the first spoken sentence. Our AI script writer picks from six angles, and they work just as well by hand:

• Problem → solution: open on the frustration the product removes, then show the fix.
• Three reasons: a fast list of why it's worth it.
• Point-of-view unboxing: first person, from getting it to using it.
• Before and after: life without it, then with it.
• Myth-busting: correct a belief your audience holds.
• Facts only: materials, warranties and specs from your page, never invented reviews or numbers.

For our demo reed diffuser, the writer chose three: “Tired of candles you can't leave burning” (example A), “Three reasons this diffuser earns its shelf space” (example B) and “From flat air to warm amber in seconds” (example C). Write three hooks for your product too and run them as separate ads: with the product and the offer unchanged, the results compare the openings. Our video ad script templates give you eight to start from.

Write a 35–50-word script

A 15–20-second video holds 35–50 spoken words: the hook, two or three facts and a call to action. Those numbers come from our own renders. Our script writer aims for 35–45 words, and the voice-overs run 11–20 seconds. Our three example videos have scripts of 44 or 45 words and run 16, 17 and 19 seconds: about 2.5–3 spoken words a second, plus just under a second of silence around the voice.

Plan at the slower rate, so a calm voice still fits. Then give each sentence one job. Here is example A, 45 words for a 17-second video:

• Hook, 10 words: “Tired of candles you can't leave burning while you're out?”
• Facts, 23 words: “This amber reed diffuser fills a medium room with warm amber, cedarwood and vanilla for about three months, no flame or plug needed.”
• How to use it, 8 words: “Just flip the reeds weekly to boost strength.”
• Call to action, 4 words: “Shop the diffuser today.”

Take every fact from your product page: no invented reviews, ratings, discounts or statistics, and no promises of results. Keep sentences short, with one idea each.

Turn your photos into shots

Plan three shots of about five seconds, each one photo with one movement, then a short end card. That fills 15–20 seconds and leaves each shot on screen long enough to read its caption.

AI image-to-video models animate a still photo. They handle camera moves well, such as a slow push-in, an orbit, a tilt or a pan, and scene details such as a light sweep, steam, water drops or moving fabric. They're far less reliable with anything the photo doesn't show: hands using the product, people, extra objects. Text and logos can warp as the camera moves, and a clip can change a detail of the product itself. Describe only camera and scene motion in each prompt, and watch every clip before you use it.

Here is the storyboard's shot plan for example A, with the caption it suggests for each shot:

1. A slow push-in on the glass bottle, with a soft light sweep across the reeds: “No flame needed”.
2. A gentle orbit around the bottle, showing the black reeds: “Safe and effortless”.
3. A slow tilt up the bottle, with warm light glowing behind it: “Warm scent, three months”.
4. The end card: “Shop the diffuser now”.

No AI video tool? A slow zoom or pan across each photo in any editor works too. It's calmer, and it can never change the product. Our photo to video page has more on choosing photos.

Add voice-over, captions, music and an end card

The voice tells the story, captions carry it without sound, music sets the pace and the end card asks for the click.

• Voice-over: record your own in a quiet room with the phone close, or use a text-to-speech voice. Aim for the pace in the table above; a rushed read is harder to follow than a shorter script.
• Captions: burn them into the video, two to four words at a time in step with the voice, in large, high-contrast type. TikTok's creative guidance gives a pace for on-screen text of “5-10 words per second” (TikTok creative best practices); captions that follow the voice run at speaking pace instead, about 2.5–3 words a second on our renders.
• Safe zone: keep captions, prices and logos clear of the app's buttons and text. For Reels ads, Meta says to keep them out of the top 14%, the bottom 35% and 6% on each side (Meta Ads Guide: Instagram Reels). TikTok publishes its safe zones only as downloadable templates, which vary by format (TikTok Ads Help: In-Feed ads, updated June 2026).
• Music: TikTok's guidance asks for sound or music. Use a track licensed for ads and keep it well under the voice. Neither TikTok nor Meta publishes a loudness target, so check the mix on a phone speaker: every word should be clear.
• End card: the last two or three seconds, with the product and one call to action of up to six words, such as “Shop the diffuser now”.

Export one file for every platform

Export a 9:16 MP4 at 1080×1920 with H.264 video and AAC audio: one file fits TikTok In-Feed ads, Instagram and Facebook Reels ads and YouTube Shorts. At 15–20 seconds it's a few megabytes (about 4–6 MB on our example renders), far below TikTok's 500 MB and Meta's 4 GB limits.

Meta's recommended 1440×2560 is the same 9:16 shape, and the minimum width for Instagram Reels ads under 30 seconds is 250 pixels, so a 1080×1920 file goes in as it is. TikTok's creative guidance asks for at least 720p. File sizes, formats and other placements are in our guides to TikTok ad specs, Reels ad specs and vertical video ad specs for every platform.

Check it and label it before you post

Watch the whole video on a phone, run through this list, and label it as AI-generated on every platform you post to.

• Specs: 9:16, 1080×1920, MP4, 15–20 seconds.
• First 3 seconds: the product on screen, and its promise both said and shown.
• Safe zone: captions, prices and the call to action clear of Meta's 14%, 35% and 6% margins, and inside TikTok's template for your ad format.
• Product: every clip still shows your product as it is, with no warped label, extra part or changed color.
• Claims: every fact matches your product page.
• Sound: the voice clear over the music, and captions that still tell the story with the sound off.
• AI label: on every platform: TikTok's AIGC label or disclaimer, “AI use” set to Yes on YouTube, a line in the ad text on Meta.

On TikTok the label isn't optional. Ads that are significantly edited or fully generated with AI need TikTok's AI-generated content (AIGC) label or a clear disclaimer, caption, watermark or sticker, and ads with undisclosed AI content can be rejected or restricted (TikTok ads policy, updated April 2026). The “AI-generated content disclaimer” in Ads Manager adds a text label at the bottom of the video (TikTok Ads Help: ad disclaimers, September 2025).

On YouTube, set “AI use” to Yes in YouTube Studio when realistic content is altered or synthetic (YouTube Help). Meta makes the disclosure mandatory for ads about social issues, elections or politics (Meta Help Center, updated around June 2026); for a product ad, say in the ad text that the video was made with AI. Our guide to labeling AI-generated ads covers the EU and FTC rules too.

The same steps in ProductToReel

ProductToReel does the hook, script, shot, sound and export steps above, starting from a product link or 1–3 photos. In about 20 seconds it writes three storyboards, each with its own hook, voice-over script, shot plan and call to action, free and without sign-up. Three credits render all three as 12–20-second 1080×1920 MP4s with AI motion from your photos, one of three AI voices, burned-in captions, music and an end card, in 3–6 minutes. The final check stays with you: our captions sit lower, and the end card's call to action higher, than Meta's Reels safe zone today, so look at the Ads Manager preview, and label the video as AI-generated wherever you post.