Merge Photos With AI: Put People and Products Together in One Believable Shot

01How merging photos with AI actually works

Combining two photos used to mean an evening in Photoshop: masking hair, matching colour, faking a shadow, and still ending up with something that looks pasted on. Image models work differently. You don't cut anything out. You hand the model the photos, tell it in plain words what you want, and it draws one new scene where the light, shadows and perspective agree.

In Vocla this happens on the Images page in the sidebar. Pick the Edit tab when you have a main photo and want to add something to it, or Create when you want a fresh scene built from pieces of several photos. Either way, you add the photos under Reference photos, and they are numbered in the order you add them: photo 1, photo 2 and so on. That numbering is the key to a clear prompt.

New to the studio? The complete design guide covers every mode; come back here for the merging details.

02Prepare your photos before you merge

The model is clever, but it builds on what you give it. A dark, cropped screenshot produces a weak merge no matter how good the prompt is. Check these before you upload:

• A clear face: front-on or a slight angle, natural light, no sunglasses or harsh shadow across it.
• The whole product: nothing cropped off the edges and the label readable, so the model can keep its shape.
• Similar camera angles: a person shot from above next to a car shot from ground level makes the perspective feel wrong.
• Simple backgrounds: the plainer the reference, the faster the model understands which object you mean.
• Enough resolution: a recent phone photo is fine. Very small images are rejected and the studio asks for a bigger one.
• A supported file: JPEG, PNG or WebP.

03Writing the merge prompt: who goes where

The number one reason a merge fails is a vague prompt like “combine these two photos”. The model doesn't know which photo is the base, where the person should stand, or what happens to the background. Use this five-part formula instead:

1. 1Name each element by number: “the person from photo 1” and “the perfume from photo 2”, not “the person” and “the product”.
2. 2Say where and how: next to, behind, holding it in her right hand, on the table in front of him, on the left of the frame.
3. 3Set the scene: a new background or which photo's background stays, plus time of day and light.
4. 4Say what must not change: “keep the face as it is” and “keep the label and colours exactly” make a real difference.
5. 5Pick the shape first: square for a feed post or 9:16 for a story, from the Shape option before you press Create.

Prompts work in English or Arabic. If you only have a rough idea, press Improve and the studio expands it into a detailed prompt in the same language, for free. Then edit it before you create.

04Ready prompts to combine two photos into one

These prompts work well as a starting point. Copy one, then swap the numbers and details for your own photos.

05Keeping faces and products intact: common failures and fixes

Quality choice matters. Vox Image runs on Google's Nano Banana 2.1 and is the studio's best quality at edits that keep faces and products the same, which makes it the default for merging. Vox Image Pro is stronger at lettering, so pick it when the merged scene needs a long headline. Even so, merges go wrong in predictable ways:

Every result opens in a viewer where you drag to compare before and after, so you see at a glance what changed. If a result is close but not quite right, don't start over: edit the result itself with a second sentence, such as “make the shadow under the car stronger”. Images that fail cost nothing; the credits come back.

06Merging for your store: from product photo to ad

The most common business use is product imagery. You have a phone shot of the product and want to see it in a model's hand, on a table in a styled room, or next to a second product for a bundle offer. Instead of booking a shoot, you upload the product and the person or place, and write a sentence. If you want the product with no background at all, run Remove background first to get a transparent PNG, then drop it into any scene.

Once you like the image, press Make it a video to turn it into an ad with an Arabic voice and captions. On current plans that starts with Starter.

07What merging costs and which plan fits

Each image is paid from your plan's credits according to the quality, and the number shows before you press Create. Current prices per image at 1K:

Rough maths: Creator at $13.99 a month gives 15,000 credits, about 37 Vox Image merges at 1K, and those credits are shared with voice and video. If you merge dozens of product shots a month or need more than 4 reference photos in a scene, Professional at $29 (80,000 credits, 10 reference photos, 4K) fits better. Prices show in your local currency and are charged in USD.

Credit rates and reference-photo limits are the current plan settings as of October 2026 and may change; the exact price always shows in the studio before you create. Nano Banana is a Google name; Vocla is an independent platform and not affiliated with Google.