Valentine's Day Couple Photo From Your Own Photos
Most couples do not have a good photograph of the two of them. They have one of her, one of him, a group shot where they are at opposite ends of a table, and a selfie taken at an awkward angle. VdoBloom's Image Editor at /dashboard/images/edit/ takes those photos and builds the picture they were supposed to have — the two of them together, in a scene you describe, with the faces coming from the uploads rather than from a written description of a person.
That last part is what separates this from a text-to-image generator. You are not describing a couple and receiving two strangers; you are supplying two real photos and telling the model how to combine them. Multi-image editing lets the instruction refer to each upload by position, so you can say what should happen to image 1 and image 2 separately, then add the constraint that matters most: keep the face, hair and identity exactly as in the source.
What this tool does
- A Valentine's Day couple photo is built on the Image Editor tab at /dashboard/images/edit/ from photos you upload, so the faces come from your pictures rather than from a text prompt.
- One edit can take several reference images at once, and the instruction addresses them by position — "image 1", "image 2" — because upload order is the order the model reads them in.
- How many references an edit accepts depends on the model: 20 on Grok Imagine Pro, 14 on Nano Banana 2 and Seedream 5 Lite, 10 on Kling IMAGE O1 and GPT Image 2, 9 on WAN 2.7 and 4 on FLUX.2 [dev].
- Identity-preserving edit models such as Nano Banana Edit and Seedream 4.5 Edit are the ones built for keeping a face recognisable, and an explicit "keep the face, hair and identity exactly as in the original" line reduces drift.
- FLUX Erase can remove an unwanted object or a third person from a mask you paint, and FLUX Outpainting extends the canvas when a photo needs to be wider.
- New accounts start with free signup credits, one-time credit packs never expire, and subscriptions are also available for a monthly allowance.
The photo you actually want is rarely the one you have
Couples accumulate photos of each other and almost none of themselves. The reason is mechanical: someone has to hold the camera. So the honest inventory is a photo of one of you, a photo of the other, and a group shot from a night out where neither face is well lit. That is enough for this workflow, and it is the whole reason it exists.
The result is a picture, not a record, and it is worth being clear with yourself about the difference. It will look like a photograph taken on a nice evening; it will not be evidence that you were there. For a card, a framed print, an anniversary post or a private joke, that distinction does not matter. For anything that will be read as a document, it does, and this is the wrong tool.
How to write an instruction that keeps both people
Instructions fail when they try to do too much. A paragraph naming the scene, the weather, the outfits, the pose, the mood and the camera angle gives the model six things to change and one thing to preserve, and the one thing it drops is usually the likeness. Write the scene in a sentence and the constraint in a sentence, and let everything else be default.
If a face still drifts, the fastest fix is almost never a new prompt. It is a better source photo: closer, sharper, more front-facing, less filtered. Editing the version you already liked with a small follow-up instruction — "keep everything, make the background softer" — preserves what worked instead of gambling on a fresh generation.
How it works
1.Pick one good photo of each of you
Front-facing, evenly lit, face clearly visible, no sunglasses and no heavy filtering. A phone photo is fine. The likeness in the result is limited by the likeness in the source, so this is the step worth taking seriously.
2.Upload both to the Image Editor
Open /dashboard/images/edit/ and add them in the same run, so the instruction can address them by position. If one of you only exists in a group photo, crop to the single person first — a face that is small in the frame gives the model much less to work with.
3.Describe the scene, then lock the faces
Two sentences. First the scene: golden-hour light, a rooftop or a quiet street, the two people from the uploads together, close and relaxed, both looking at the camera. Then the constraint: keep each face, hair and identity exactly as in the source photos. The constraint is not optional.
4.Check both faces at full size, then correct
Look at each face separately before you look at the picture as a whole — the whole is what makes you save it, the faces are what make it right. If one drifted, swap that source photo or add a follow-up instruction on the version you liked rather than generating again from the start.
Frequently asked questions
How do I make a Valentine's Day couple photo if we have no photo together?
Upload one photo of each of you to the Image Editor at /dashboard/images/edit/ and describe the scene you want, naming the uploads by position. The faces come from those photos, so the result looks like the two of you rather than like a pair of generated people. It solves the missing-photo problem instead of the photography problem.
Will it look like us, or like similar-looking strangers?
The faces come from your uploads, and the identity-preserving edit models — Nano Banana Edit and Seedream 4.5 Edit — are the ones built for keeping a face recognisable. Adding "keep the face, hair and identity exactly as in the original" to the instruction measurably reduces drift. Locking the faces every time is the habit that keeps a made photo believable.
What photos work best as inputs?
One sharp, front-facing photo per person with the face filling a good part of the frame and even lighting. Avoid sunglasses, heavy filters, strong side shadows and shots where the person is small in a crowd. A plain, well-lit phone portrait beats a beautiful photo where the face is turned away or in shadow.
Can I put the two of us in a specific setting?
Yes, and this is where the tool is more useful than a filter — the scene is whatever you describe. A restaurant, a beach at sunset, a street you both know, a plain studio backdrop with warm light. Describe the setting and the framing in one sentence, and keep the description of the people to the constraint line.
Can I remove a third person from a photo so it is just us?
Yes. Select FLUX Erase, paint a mask over the person you want gone, and generate — the model fills the masked area from the surrounding image. You can also just instruct an edit model to remove the person on the left and fill the background naturally. It is the usual fix when the only photo of the two of you has someone else in it.
Is the result a photo or a video?
This page is the still image. The same two photos can be turned into a short video instead — the video side of the site takes a photo of a couple and animates the scene around them, and the kissing effect takes either one photo with two people or two separate photos and preserves the facial features and identity from the originals.
Do I need a subscription to make one couple photo?
No. The tool runs on credits rather than a plan, and new accounts start with free signup credits, so a single Valentine's photo does not require subscribing. One-time credit packs never expire, which fits a picture or two a year; subscriptions are there for people using the tools regularly.
Related tools
AI Christmas Couple Photo: Two Faces, One Frame
Make an AI Christmas couple photo from one picture. Choose the Christmas Couple preset and get a festive still back — both faces kept from your upload.
AI Image Editor — Edit Any Photo By Typing What You Want
Edit any photo by typing an instruction. 30+ models incl. Nano Banana 2, Seedream 5, Flux Kontext, WAN 2.7 — mask erase, outpaint, up to 4K output.
AI Fashion Try-On — See Any Outfit On Yourself Before You Buy
Upload your photo plus an outfit photo. AI fits the exact garment on you in 6 looks: studio catalog, street, outdoor, party, mirror selfie, editorial.
Reverse-Engineer a Reusable AI Prompt From Any Photo
Turn any image into a structured AI prompt — pose, outfit, hair, lighting, camera and constraints, in a fixed section order. Send it straight to Edit.
Ready to try it?
Make a couple photo