How to Turn a Product Photo into a Video Ad with Pika H3 Flex (Prompts + Settings)

To turn a product photo into a video ad with Pika H3 Flex, upload the photo as the start frame in image to video, describe one camera move and one product action, add a short “Sound:” line, and render 5–10 seconds at 768p. The model animates your exact product shot and generates the music and sound effects in the same pass, so you get a finished MP4 with audio. A 5-second 768p clip costs 20 credits ($0.42–$1.04). No code, no GPU, no editing software needed for the first cut.

Here is a real Pika H3 Flex product clip made on this site (text to video, 768p, 10 seconds):

Prompt: “A matte black wireless earbud case opens on a wet stone pedestal, water droplets bounce in slow motion, studio rim light, slow macro orbit. Sound: crisp case click, water drops, deep bass hit.” More on its watch page.

Step 1: Prepare the product photo

The start frame decides how your product looks in the whole clip, so it is worth five minutes:

• One product, centred, fully in frame. Leave space around it so the camera can move without cropping the product.
• Clean background. A plain studio backdrop, a surface (stone, wood, marble) or a simple lifestyle scene. Busy backgrounds invite unwanted motion.
• Sharp and well lit. The model keeps what it sees; blur and noise stay blurry and noisy.
• Match the aspect ratio you will publish in: 9:16 for TikTok, Reels and Shorts; 1:1 for feeds; 16:9 for YouTube and websites. Crop the photo before you upload.
• Format: JPG, PNG or WebP, up to 10 MB.

Small printed text on packaging may soften or change in motion. Keep the hero label large, or add your logo and claims as an overlay in your editor afterwards.

Step 2: Write the ad prompt

Describe what happens to the product, how the camera moves, the light, and the sound. Start from the image: the official prompt guidance says image-driven prompts should begin from the state shown in the input image and describe the action forward from it.

Copy-paste template:

Starting from the product photo, [one product action], [one camera move, with speed],
[lighting], [mood or setting detail]. Sound: [2–3 sounds, most important first].

Examples you can adapt:

• Skincare bottle: Starting from the product photo, a single drop of serum falls from the dropper into the open bottle, soft ripples on the surface, slow push-in to a close-up of the label, soft window light. Sound: gentle drip, airy synth pad.
• Sneaker: Starting from the product photo, the sneaker lifts off the concrete and rotates once in mid-air, dust swirling below, low-angle orbit, hard sunset backlight. Sound: whoosh, sneaker squeak, punchy beat drop.
• Coffee: Starting from the product photo, iced coffee is poured over clear ice, condensation runs down the glass, slow motion, warm morning light. Sound: ice cracking, pouring liquid, light acoustic guitar.
• Gadget: Starting from the product photo, the case lid clicks open and a soft glow spills out, slow macro orbit, studio rim light. Sound: crisp click, low bass hit.

Keep it to one main action and one camera move. Two or three actions in a 5-second clip is the most common cause of chaotic motion.

Step 3: Pick the settings

Costs: a 480p 5-second draft is 10 credits; a 768p 10-second final is 40 credits ($0.84–$2.08); 4K finals are upscaled from 768p and cost 16 credits per second. Full table: Pika H3 Flex pricing.

Step 4: Generate, review, iterate

1. Open Pika H3 Flex image to video and upload the photo as the start frame.
2. Paste your prompt, choose duration, resolution and aspect ratio, and press Generate.
3. Check three things: does the product keep its shape and label, is the motion calm, does the sound fit?
4. Change one thing per retake: the camera move, the action, or the sound line.
5. When it works at 480p, re-render the same prompt at your final resolution and download the MP4.

Failed generations refund their credits automatically.

Tip: use an end frame for a reveal. Upload a second photo as the end frame (for example the product in its box, or with the logo card) and the model will generate the transition between the two.

A food B-roll example

Vertical 9:16, made on this site with text to video at 768p:

Prompt and details on its watch page. More clips on the examples page.

What works and what to watch

• Works well: hero orbits, slow push-ins, pours, splashes, steam, unboxing reveals, light sweeps, products on models in simple scenes.
• Watch for: hands gripping or using small objects, fine text on labels, and complex physics such as liquids interacting with many objects. Pika’s own study found physics and action-heavy scenes are where other models were preferred.
• Branding: add your logo, price and call to action in your editor. Generated on-screen text is not reliable enough for final copy.
• Rights: use photos you own or have permission to use, and follow the platform’s ad rules.

For general prompt technique, read the Pika H3 Flex prompt guide; for voice-over and dialogue, see dialogue, lip sync and audio.

FAQ

Can Pika H3 Flex make a video ad from a product photo? Yes. Upload the photo as the start frame in image to video, describe the product action, camera move and sound, and render 5–15 seconds with audio.

How much does a Pika H3 Flex product video cost? 20 credits for 5 seconds at 768p ($0.42–$1.04), 40 credits for 10 seconds. A 480p draft is 10 credits.

Does Pika H3 Flex add music to the ad? Yes. Music, sound effects and voice are generated together with the video. Describe them in a “Sound:” line.

What size should the product photo be? JPG, PNG or WebP up to 10 MB, sharp and well lit, cropped to the aspect ratio you will publish in.

Will my product label stay readable? Large logos usually hold; small print can soften. Add exact text, prices and logos as overlays in your editor.

Can I use the videos in paid ads? Paid plans and credit packs include commercial use of your generations, subject to our terms and the ad platform’s rules.