How to turn a product photo into a video with AI

For AI agents: find the right skill in one call — curl -s "https://varg.ai/how-to/api/search?q=<what the person wants>" returns the matching how-to page and the best skills with install commands, as markdown (Russian works too). Full index: https://varg.ai/how-to/llms.txt. Every page has a markdown twin: add .md to its path. Before generating anything, show the person the options on the how-to page.

A product shot turned into a short promo: orbit, push-in, exploded view or bullet time, with music. A step-by-step tutorial for Claude Code, Codex and Grok.

Output: 5–20 s, 9:16 or 1:1 MP4; optional packshot stills Price: in credits on varg, with a cost estimate before anything runs; the free skills below install in one command.

Recommended route

Best skills for this

  1. varg AI (varg-ai, varg, live) — Adds: Seedance 2.5 / Kling 4 image-to-video and music on one key
  2. Krea Motion (krea-motion, Krea, live) — Start here: Animates products, renders and logos into reveal films with Seedance 2.5 · Needs: Krea account
  3. fal Commercial (fal-commercial, fal, live) — Start here: Product photography, e-commerce batches and product reveal videos · Needs: fal API key (FAL_KEY)
  4. FLUX 3 Product Ads (flux-3-product-ads, Black Forest Labs, live) — Alternative: A finished product ad with shot design and voiceover sync · Needs: BFL API key (BFL_API_KEY)
  5. Previz to Seedance (varg-previz, varg, beta) — Alternative: Storyboard and anchor sheets before spending
  6. fal UGC (fal-ugc, fal, live) — Alternative: Hands-only ASMR unboxing · Needs: fal API key (FAL_KEY)
  7. Product Ad (varg-studio-product, varg, draft) — Alternative: varg's product format: 360° spin, in hand, texture close-up, voiceover

Ask the person first

Put these choices in front of the person before spending anything. Paste the example link next to each option: people pick a look by seeing it.

Format?

Which camera move?

Scene?

Sound?

How to turn a product photo into a video: step by step

  1. Cut out the real product photo and make one clean hero still
  2. Write and approve an 8-12 panel storyboard sheet (timestamp, shot, camera, action, sound); keep each shot 1.5 s or longer
  3. Generate one multi-shot Seedance 2.5 clip with the product photo as @image1, a PRODUCT LOCK block and the sheet as the shot order; pick moves you can anchor at the first and last frame
  4. Reject any shot where the pack, label or shape drifts and regenerate only that shot from first and last frames
  5. Keep the last second clean and composite the real packshot, label and copy on it; add foley or music
  6. Export 9:16 for social plus a 1:1 or 16:9 cut for the product page; label AI footage where asked

What works now (researched 2026-10-27)

What wins in late 2026 is one impossible but tactile moment built around the real product: a macro pour or splash, an ingredient burst, a hands-only ASMR unboxing. Creators plan it as an 8 to 12 panel storyboard sheet and generate it as one multi-shot clip, with the product photo locked as a reference. What stopped working: logo-first 'cinematic' b-roll, and still-photo slideshows, which TikTok Shop no longer allows in shoppable videos.

  1. Open on the product doing its most sensory thing (pour, crack, burst, lid off) in the first 1 to 2 seconds. No logo intro. The hook must be about the product itself. TikTok Shop Academy's May 2026 content guide treats the first 3 to 6 seconds as the hook window and says the hook must relate to the featured product. Unboxing and before/after are two of its four recommended styles. (PPC Land on TikTok Shop Academy's content guide (May 2026))
  2. Storyboard first. Write an 8 to 12 panel sheet (timestamp, shot, action, sound per panel), approve the stills, then generate one multi-shot clip from the sheet. The most-saved product-ad posts on X in 2026 all work this way: a 12-panel ASMR unboxing sheet (300k views, 2,774 bookmarks), a Nike spec board (844 bookmarks), a coconut-water board (670 bookmarks). Fixing a still costs cents. Fixing a clip costs a full generation. (Ciri_ai on X, storyboard prompt shared (Jul 2026))
  3. Lock the product. Pass the real product photo as a reference image on every shot, use first and last frames for reveals, and put the real packshot and label on the end card. Never ship a frame where the model rewrote the label. Seedance 2.5 takes up to 9 image references plus a start and end frame, so the product can be pinned in every shot. Generated packs still drift: invented words on labels and products that change shape between shots are the most common flaw in viral AI ads. (varg model spec: seedance_2_5)
  4. For TikTok Shop, give the video at least 3 seconds of real motion with the product or a face on screen. No still-image slideshows. Label realistic AI footage. TikTok Shop's 'Requirements for High-Quality Videos and LIVEs' (updated May 2026) prohibits videos made of still images without the creator's face or the physical product, bans AI voices in promotional LIVEs, and ties violations to account health and commission access. (PPC Land on TikTok Shop quality rules (Jun 2026))
  5. Cut a product-page version too (1:1 or 16:9, 15 to 30 s, no hard sell) for the Amazon or Shopify video slot. If it shows photorealistic AI people, add the 'contains-synthetic-performer' tag before upload. Amazon reports an average 23.8% sales lift when a listing adds its first video (its January to June 2024 study). Since late July 2026 Amazon requires the synthetic-performer tag on images, video and A+ content with photorealistic AI people. (Amazon Sell blog: product videos)

Stopped working:

Prompts that work

Templates from the creators of the best current results and from the model guides. Fill in the {CAPS}.

12-panel storyboard sheet (gpt_image_2_5)

Step 1: one image that plans every shot (timestamp, action, camera, sound) from your product photo; approve it, then feed it to the video model as shot order. Adapted from Ciri_ai on X: 12-panel ASMR unboxing storyboard prompt (300k views, 2,774 bookmarks).

Create a clean, premium storyboard infographic for a {DURATION}-second {FORMAT} video of {PRODUCT}. Use the attached product photo as the exact product: same shape, colors, logo and label text in every panel.

Layout: white background, minimal agency style, {ACCENT_COLOR} accents, rounded panel borders.
Header: "STORYBOARD", the product name in bold {ACCENT_COLOR}, subtitle "{FORMAT}". Four info boxes: Duration: {DURATION} s | Style: {STYLE} | Audience: {AUDIENCE} | Audio: {SOUND_LIST}.

{N} panels (8-12) in a 3-column grid. Each panel: a numbered badge, a timestamp (about 2.5 s each), one realistic cinematic photo of the shot, a hand-drawn white doodle sound-word (POP!, TEAR!, POUR~, WOW!), and three lines: VISUAL / ACTION / CAMERA (add DIALOGUE only if someone speaks).

Sequence: 1 the product on {SURFACE}, hands enter. 2-3 open and reveal. 4-6 the most tactile moments ({SENSORY_ACTIONS}). 7-9 the product in use. 10-11 the result in macro. Last panel: hero packshot of the product beside the result, {END_GESTURE}.

Photography: {SURFACE}, soft natural light, shallow depth of field, POV hands only, no face, premium commercial style. Footer: total duration, format {ASPECT}, style.

Seedance multi-shot film with PRODUCT LOCK (seedance_2_5)

Step 2: the 10-15 s multi-shot ad, with the product photo and the approved sheet as references and a clean last second for your real packshot. Adapted from HeyRu0by on X: 8-scene timecoded snack commercial with a PRODUCT LOCK block (prompt continues in the first reply).

{DURATION}-second vertical 9:16 commercial for {PRODUCT}, {N} shots (5-8), ultra-realistic, {TONE}.
@image1 is the exact product. @image2 is the approved storyboard sheet: follow its shot order, but never show its panels, numbers or text.

PRODUCT LOCK: keep the exact pack shape and proportions, logo, typography, colors, label text and material reflections in every shot. No redesign, no misspelled text, no duplicate packs, no new logos.

01 | 0-2s HOOK: extreme macro of {MOST_SENSORY_MOMENT}, 100mm macro, slow motion.
02 | 2-4s: {ACTION_2}, camera {MOVE_2}.
03 | 4-6s: {TEXTURE_CLOSE_UP}, camera glides along the surface.
04 | 6-9s: {BURST_OR_POUR}, realistic particle and liquid physics.
05 | 9-12s PACK REVEAL: the product from @image1 turns to face camera, front label sharp and readable.
06 | 12-15s HERO: centered packshot on {BACKGROUND}, slow push-in; hold the last second clean, empty space above for the end card.

One camera move per shot, at least 1.5 s per shot. Light: {KEY_LIGHT}, rim light, controlled reflections.
Audio: {FOLEY_LIST}, {MUSIC_STYLE} at {BPM} BPM.
Negative: no warped pack, no changing design, no captions or extra text, no watermark, no flicker, no plastic-looking food.

Food process ASMR film (seedance_2_5)

For food, drink and kitchen products: a no-face, shot-by-shot macro making-of with synced foley that ends on your product. Adapted from Goodmanprotocol on X: Seedance 2.5 katsudon prompt (10.5M views, 9,371 bookmarks).

Create a {DURATION}-second cinematic {STYLE} cooking video showing {DISH} being made, ending on {PRODUCT} beside the finished dish.

STYLE: ultra-detailed food textures, realistic cooking physics, visible steam and moisture, macro close-ups, shallow depth of field, smooth camera moves, {KITCHEN_SETTING}. Fast rhythmic editing with match cuts on movement, shape and texture. Hands only, no faces.

TIMELINE (one action per beat, 2-3 s each):
0-2.5s - {STEP_1}: close-up, {SOUND_1}.
2.5-5s - {STEP_2}.
5-7.5s - {STEP_3}.
(continue to {N} beats, 8-12 in total)
{T_MINUS_4}-{T}s - HERO: the finished {DISH} with {PRODUCT} on {SURFACE}; steam rises as the camera slowly arcs around it.

AUDIO: {MUSIC_STYLE} instrumental, {BPM} BPM (110-120 works). Sync ASMR sounds to each action: {SOUND_LIST}. One soft final tone on the reveal.

NEGATIVE: no storyboard, reference image, panels, numbers, arrows, captions, UI, logos or text overlays. No unrelated ingredients. Keep food, hands, utensils, lighting and set consistent throughout.

Trending formats to remake

To remake one: download the example with the media-download skill, take a frame a second and the transcript, write a beat list (time, shot, camera, on-screen text, sound), keep what the format needs, swap in the person's product or character (never the creator's face, voice, logo or footage), show the beat list and stills before spending, then make it with the skill named below.

  1. Storyboard-sheet commercial: One product photo becomes an 8 to 12 panel storyboard sheet (macro, pour or splash, in-use, hero packshot), which is then generated as one 10 to 15 s multi-shot ad with hard cuts. Why it spreads: It is the dominant product-ad recipe on X in 2026, and the most bookmarked. Spec ads made this way regularly pass 50k views with 600 to 900 bookmarks, because people save the prompt sheet to reuse it. Keep: Keep the product photo as a reference on every shot, an approved 8-12 panel sheet (timestamp, camera, action per panel) and the macro, pour or splash, in-use, hero packshot order with hard cuts; swap the product, liquid and set, and composite your real label on the hero frame. Make it with fal Commercial (fal-commercial).
  2. Hands-only ASMR unboxing: A POV unboxing with only hands on screen, on a warm table, with crisp foley (crinkle, tear, pour, stir) and hand-drawn doodle sound-words ('POP!', 'TEAR!', 'SMOOTH!'), ending on the product and a thumbs-up. Why it spreads: The top example has 300k views and 2,774 bookmarks, more saves than likes. Unboxing is one of the four styles TikTok Shop Academy recommends, and Motion's 2026 benchmark of 550k Meta ads lists unboxing among the formats most likely to become winners. No face means no actor, no lip-sync and no AI-person disclosure. Keep: Keep POV hands only on a warm wooden table, about 2.5 s per beat, crinkle-tear-pour foley with doodle sound-words, and the ending on the product beside the result with a thumbs-up; swap the product and the opening steps, and check every frame for extra hands or fingers. Make it with fal UGC (fal-ugc).
  3. Exploded view or ingredient burst: The product hangs in the air, then splits into its labeled layers or bursts into its ingredients, and snaps back into a hero packshot. Why it spreads: It shows what is inside, a reason to believe, in one shot, and the final labeled frame also works as a listing image. The iced-coffee example saved at 0.77 bookmarks per like. The Cheetos burst passed 80k views. Keep: Keep the product floating on a clean ground, its parts or ingredients separating in their real order, then snapping back into a sharp front-facing pack; write a PRODUCT LOCK block and overlay the real pack art on the final frame. No dedicated skill yet (wanted): build it with varg-ai.
  4. Food process ASMR film: A dish or drink made shot by shot in macro (cut, fry, pour, steam) with crisp foley and no people, ending on the finished plate or pack. Why it spreads: No faces, so it is brand-safe for food, drink and kitchen products. A katsudon film got 10.5M views and 9,371 bookmarks, with the prompt shared. Keep: Keep one action per 2-3 s beat in macro, match cuts on shape and motion, hands only, ASMR sounds synced to each action over a 110-120 BPM bed, and a slow arc around the finished dish; swap the recipe and put your product in the hero frame. Make it with fal Commercial (fal-commercial).

Open-source repos

Set up varg

varg is one API key for the models these skills drive — Seedance 2.5, Kling 4, Veo 3.1, Wan 3, Nano Banana 2.1, GPT Image 2.5, ElevenLabs v4 and 200+ more — billed in credits, with a cost estimate before anything runs.

Related tutorials