authustle
Step by step

How to put yourself in any scene

Put yourself in a scene works from photographs and a sentence: several pictures of you go in as identity references, you describe the shot you want in plain words, and the model builds it with you inside. There is no source video and nothing to copy — this is the one tool where the scene is yours to invent.

Open the Studio

No credit card to sign up · 5 steps

Three fields, and the middle one is the work

  1. 1

    Open Studio and pick Put yourself in a scene

    The tool is described as your photos plus a described scene. Opening it shows the form and names the reference-to-video model behind it.

    studio/me-in-scene

  2. 2

    Upload references under Your photos

    The first field is labelled Your photos and takes between one and nine. More angles and more lighting conditions give the model more to hold on to; nine near-identical selfies from the same afternoon are worth about as much as one. The form caps the count at what the model actually accepts.

    studio/me-in-scene#your-photos

  3. 3

    Describe the shot under The scene

    The second field is labelled The scene, takes at least 3 and up to 800 characters, and its placeholder shows the register: "Walking through a neon-lit Tokyo street at night, handheld camera following from the front". Name the action, the setting and the camera. Adjectives about mood are the least useful thing you can spend those characters on.

    studio/me-in-scene#the-scene

  4. 4

    Set the length and press Generate

    The duration slider is bounded by the model's own limits. The form prints the exact credit cost of the run next to the Generate button, and re-prints it whenever you change the length. Nothing is charged before you press it.

    studio/me-in-scene#duration

  5. 5

    Collect the clip from your Library

    The finished vertical MP4 lands in your Library. Runs are processed in the background, so you can close the tab — the finished file turns up in your Library. We do not quote a fixed wall-clock time: it moves with the length of the run and with how busy the queue is.

    library

Which reference-to-video model to use

HappyHorse 1.1 Reference to Video

Product default

The default, and the catalogue's recommended pick for this job. It is what the tool runs on and what the nine-photo cap is sized around.

The trade-off: Identity-first models spend their capacity on keeping you recognisable, which is not the same as spending it on the scenery behind you.

Specs and credit price →

Kling O3 Pro Reference to Video

From the model spec

The scene itself is the point — a location, a camera move, a piece of staging you want rendered convincingly around you.

The trade-off: A different price and a different reference-image ceiling, both printed on its catalogue page.

Specs and credit price →

MiniMax H3 Reference to Video

From the model spec

A third family to try when a scene comes back nearly right and you want a different set of failure modes on the same prompt.

The trade-off: Different failure modes are not fewer failure modes — check its own page for what it accepts before you re-run.

Specs and credit price →

These picks come from our own side-by-side runs and from what the tools actually default to. The specs and credit prices live on each model's own page and are generated from the catalogue, so they cannot drift out of date here. We do not quote anyone else's review, benchmark or price.

What it costs to follow this guide

Your first video from $0.99

A one-time purchase — 35 credits land in your balance immediately. No subscription, nothing renews, and if a generation fails on our side the credits come back automatically.

What this does not do

  • It does not copy a video. There is no source clip here — if you want an existing Reel's pacing, that is the autopilot.
  • It does not put a real, identifiable other person into your scene. Upload your own likeness, not someone else's.
  • It does not read a long prompt as a shot list. One scene, one shot; 800 characters is plenty and usually too many.
  • It does not post anywhere. We never ask for your Instagram, TikTok or YouTube password, never read your DMs and never publish anything for you. You download the file and post it yourself.
FAQ

Common questions

How many reference photos should I upload?
The form takes up to nine and one is enough to run. Variety is what pays: different angles, different lighting, different days. Nine frames from one burst add almost nothing over the first of them.
How should I write the scene?
Like the placeholder does — the action, the setting and the camera in one sentence. "Walking through a neon-lit Tokyo street at night, handheld camera following from the front" is a better brief than three lines about atmosphere.
Why does my face change between runs?
Reference models rebuild you each time from the photos you gave them, so weak or repetitive references show up as drift. More varied references is the fix that actually works.
What does it cost to try?
Your first video from $0.99. A one-time purchase — 35 credits land in your balance immediately. No subscription, nothing renews, and if a generation fails on our side the credits come back automatically.

Keep reading

Put yourself somewhere else

Bring a few photos and one sentence. The run is quoted in credits before you press Generate.

Open the Studio

No credit card to sign up