How to Do the 'Generate a Man Next to Me That Would Suit Me' AI Trend (2026)

What You're Trying to Make

You upload one selfie, and AI casts the partner it thinks would suit you — then, instead of a single frozen photo, it plays a candid handheld home-movie of your whole life together: a lazy morning sharing coffee, goofing around cooking, getting dressed up for a date night, a candlelit dinner, a sunny getaway, and a quiet night in on the couch. It looks like someone found the tape of your perfect relationship. You never met this person; the whole thing is generated from one photo of you.

This is the video upgrade of the viral "generate a man next to me that would suit me" trend. The original is a still image — a partner dropped into the empty space beside you. This guide covers the whole thing: where the trend came from, the exact prompt recipe to make the moving version, and the one-tap way to make yours.

Fastest way — The One Who'd Suit Me on Starrd casts your ideal partner and renders the full candid montage from one selfie — first video free, no prompt to write. Want the full method first? Read on. ↓

Is This a Trend? (Yes — and It Has a Long Lineage)

It's real, and it's been building for years. It started as the 2022–23 "soulmate sketch" craze — static AI or psychic-style portraits of a supposed future partner. It leveled up in spring 2025, when chatbots got good enough at native image generation to take your own selfie and imagine an ideal partner standing beside you — which spawned the running joke that these models default to a suspiciously consistent "type." By January 2026 it crystallized on TikTok as the copy-paste line — "generate a man next to me that would actually suit me" (also phrased "…that you think would suit me"): post a photo with empty space beside you, ask the AI to add a partner without changing your face, and compare the result to your real one — all set to the trend's now-signature audio.

Why it keeps working: it's a personality test disguised as a magic trick. The result is about you — your vibe, your "type," whether the AI "gets" you — so everyone wants to see their own. Every wave to date has been a single still image, which is exactly the opening: nobody has made the moving version where you don't just meet your ideal partner, you watch a whole life with them play out. That's the white space this format fills.

The Fastest Way — Use "The One Who'd Suit Me" Template on Starrd

The template is live in the Starrd library. It packages every step in this guide — casting a believable ideal partner, keeping your face, locking the couple so they stay the same two people across every scene, and the nostalgic home-video grade — into a single upload.

1. Pick one clear selfie. Front or three-quarter view, eyes open, good lighting.
2. Open "The One Who'd Suit Me" in the Starrd app or web library.
3. Upload and tap generate. It casts your ideal partner, builds a six-scene storyboard of your life together, and renders a candid vertical montage on Seedance 2.0 — audio included.

One selfie, a few minutes. Want a specific type? Type it in Director's Notes:

• Your type — tall with curly hair, glasses, into books, athletic, tattoos; the partner is built to match. Leave it blank and the AI casts someone that complements you.
• The vibe — cozy and wholesome by default; you can lean it more playful or more romantic.
• The scenes — swap the getaway (beach, city, road trip) or a beat that isn't your thing.

The rest of this guide is for people who want to roll their own — control the partner, the scenes, or run it on a different model.

Or, Build It Yourself — What You Need

Three things:

1. One clear selfie of yourself. Front or three-quarter view, eyes open, good lighting.
2. An image model to cast the couple + a video model to move them. The trick is to generate a storyboard of the whole couple first (GPT Image 2 is great at keeping two people consistent), then animate it with a model that takes a reference image — Seedance 2.0 is the safest.
3. A platform to post on. TikTok, Reels, and Shorts are where this travels.

You don't need a partner, a film crew, or a single second of real footage.

Step 1 — Pick Your Reference Photo

The selfie you feed the model is the face that ends up in the video. Choosing well saves wasted generations.

Use:

• A clear, well-lit photo of just you
• Front-facing or three-quarter angle, eyes open
• A head-and-shoulders or half-body shot (the model handles the styling — you supply the face)

Avoid:

• Group photos (the model gets confused about who the "you" is)
• Sunglasses or anything covering the face
• Low-resolution or motion-blurred shots
• AI-generated reference images (compounding artifacts)

Step 2 — Lock the Concept

Decide the details before you write the prompt:

• 🧑‍🤝‍🧑 The pairing: you (from your selfie) plus an invented ideal partner — one believable, attractive, everyday person, not a model-perfect mannequin and never a real celebrity.
• 💬 Their type: let the AI cast someone that complements you, or specify it (hair, height, style, glasses, tattoos).
• 🎬 The scenes: a life-together montage beats a single moment. A good arc: cozy morning → cooking together → date-night prep → a candlelit dinner → a getaway → a night in. Vary the emotion so it doesn't repeat.
• 📼 The look: nostalgic early-2000s camcorder — faded colors, a little grain, gentle handheld — is what makes it feel like a real found home video instead of an AI render.

Step 3 — Write the Prompt

The secret to this one is a storyboard first: generate all six couple scenes as a single image so the two people are cast once and stay consistent, then animate that storyboard into a montage. Copy this and swap in your specifics:

STORYBOARD (one image, six panels): the same happy couple across six candid scenes of a life together. Person A is the subject from the uploaded selfie — keep their face exactly. Person B is an invented ideal partner: one consistent, genuinely attractive, everyday adult, kept the exact same face and hair in every panel (not a celebrity, not a mannequin). 1) a lazy morning sharing coffee in a sunlit kitchen, 2) cooking together and laughing, 3) getting ready for date night at a mirror, 4) a candlelit dinner toast, 5) a sunny travel getaway overlooking the coast, 6) a cozy night in on the couch. Candid and un-posed — they look at each other, never at the camera. Small corner numbers only, no captions.VIDEO (12s, vertical 9:16): read the six panels in order and animate them as ONE warm early-2000s home-video montage with gentle cuts between the moments. Early-2000s DV camcorder look — soft faded colors, mild grain, gentle handheld drift, occasional autofocus hunt, no modern color grading. A friend is loosely filming; the couple is absorbed in each other and never poses for the lens. Keep both faces identical across every scene; only the outfits change per scene. Calm, settled motion — no fast or jerky movement. Natural ambient audio only (kitchen sounds, easy laughter, a breeze), no music track so a trending sound can be added.

The non-negotiable elements:

• Candid, not posed — the couple absorbed in each other, never mugging at the lens. Selfies and to-camera smiles read as staged.
• One consistent partner — cast them once in a single storyboard so they don't morph scene to scene.
• Your face, locked — the whole trend is you plus the partner. Keep your selfie's likeness exact.
• Calm camerawork — gentle handheld, not frantic. Heavy shake plus lots of walking reads as choppy.

Step 4 — Pick a Model

• Seedance 2.0 — what Starrd runs on. Takes the storyboard + your selfie as reference images, holds both faces across the montage, and generates the ambient audio in the same pass. Safest pick.
• GPT Image 2 — for the storyboard step; it's unusually good at keeping two people consistent across panels.
• Kling / Veo 3.1 — can animate the storyboard too; push "candid, un-posed, no looking at camera" hard or they drift toward posed shots.

The static-photo version of the trend is just the first half — a chatbot (ChatGPT or Gemini) adding a partner to your photo. The video is what you get when you storyboard that couple and animate it. No preference? Start with Seedance 2.0. (See our Seedance 2.0 Prompt Guide.)

Step 5 — Generate and Iterate (Common Failure Modes)

First generations rarely nail it. The usual failures and fixes:

The partner looks like a different person in each scene. You generated the scenes separately. Build them as one storyboard image so the partner is cast once, then animate that — the single generation is what locks the face.

It feels staged, like a photoshoot. Too many to-camera smiles and selfie poses. Recompose every beat as observed — the couple looking at each other or at what they're doing, with a friend loosely filming. Add "candid, un-posed, never looking at the camera."

Your own face drifts. Use a clearer, front-facing selfie, and pass it as a reference alongside the storyboard so the model has a crisp copy of your likeness to hold onto.

The motion looks choppy / low-framerate. Usually heavy camera shake stacked on top of lots of walking. Keep the beats calm and mostly stationary (sharing coffee, sitting, cuddling) and the handheld gentle — the nostalgic look comes from the grade, not frantic motion.

The partner looks fake or model-perfect. Ask for a "believable, attractive everyday person, not a plastic mannequin." Real-looking beats runway-perfect for this format.

Budget 2–4 generations before a keeper.

Step 6 — Post It

Lean into the reveal. Caption it like a reveal — "AI picked the one who'd suit me 🥹" — or play up a funny miss. The AI-cast angle is the hook (and it doubles as your AI-disclosure).

Add the trending sound. The prompt says "no music" so you can lay the trend's audio over the top where this format lives.

Vertical, and let it breathe. It's a 9:16 montage — the quiet, warm moments are the point, so don't over-cut it.

Common Mistakes That Tank Your Video

1. Casting the partner six times. Separate generations = six different faces. One storyboard = one consistent person.
2. Posed, to-camera shots. The magic is candid. The second they smile at the lens, it stops looking like a real home video.
3. A weak reference selfie. Blurry, dark, or group shots = your own face drifts. Spend your effort here.
4. A model-perfect partner. Uncanny, glossy "AI model" looks kill the believability. Ask for an everyday person.
5. Treating it as real matchmaking. It's for fun. Frame it as entertainment, not a prediction — that's the tone that travels.

Frequently Asked Questions

What is the "generate a man next to me" trend? Upload a photo of yourself and AI adds a partner it thinks would suit you. The video version plays a candid home-movie of your whole life together instead of a single still.

How do I make one? Storyboard the couple in one image (you + an invented ideal partner), then animate it on a reference-image model like Seedance 2.0 — or use "The One Who'd Suit Me" template and just upload one selfie.

Does it work for a girlfriend too? Yes — it's sex-agnostic. Specify your type in Director's Notes, or let the AI cast a fitting partner.

Can AI actually tell who'd suit me? No — it's entertainment, not matchmaking. The invented partner is for a fun, shareable clip.

Do I need to label it as AI? Yes — all major platforms require it, and the AI-cast angle is the whole joke anyway.

Related Reading

• How to Make a Red Carpet AI Video — the other self-insert flex: you stepping out of an SUV into a wall of paparazzi flashes (solo or a couple).
• How to Make a Day in the Life (Korea) AI Video — the nostalgic 2000s-camcorder home-video look behind this montage.
• How to Make a Skyfall AI Video — the longest-running self-insert trend of the year.
• Viral AI Video Trends (2026): The Monthly Roundup — every trend worth making this month, each with a one-tap template.
• Seedance 2.0 Prompt Guide — the framework behind the storyboard-plus-montage prompt above.