After Dark
Laser Room
Color Smoke
Silk Drop
Petal Storm
Slow Shutter
Candlelight
Behind the Hair
Call Me
Champagne Toast
Darkroom Red
Doorway Lean
Eyes on You
Join Me
Steamy Mirror
Late-Night Dance
Club StrobeRomance
KissCamera moves
Tiny PlanetTransform
Molten GoldFashion
Mirror GlanceAtmosphere
Lace ShadowsMotion
GrooveFun
MeltdownProduct
Ingredients BurstFilm looks
Film BurnVideo templates
The Glance Back
Candles Come Alive
Silk Drift
Neon After Hours
Sway Under the Lamp
Mirror Moment
Morning Light
Passing Headlights
Whisper Session
Slow Push-In
Half Orbit
Drone RiseThis effect writes the prompt for you. Add one image.
We never filter your prompt.
Steam fogs the glass between the camera and your subject until a hand wipes a clear streak and reveals a knowing smile.
- Model
- MiniMax H3
- Length
- 5 s
- Resolution
- 768p
- Per run
- $0.185
- Type
- After Dark
- Also runs on
- Seedance 2.0 Fast, Seedance 2.0, Wan 3.0, Vidu Q3 Turbo, Wan 2.6 Flash, Seedance 2.0 Mini
What this effect sends
This is the full prompt, word for word. Copy it, or open it in the generator and change anything you like.
The camera is a mirror. Within the first second thick steam fogs the glass between the lens and the person until they are only a warm blur. By the second second they reach toward the lens and wipe one wide clear streak across the fogged glass with the flat of their hand, and through the clean strip their face appears sharp, looking straight into the lens with a knowing smile. Fresh fog creeps back in at the edges of the streak while the camera holds still. The person stays exactly as they appear in the photo.Pictures that work best
- Choose a photo with the face clearly visible and looking at or near the camera.
- Keep a hand or arm in the photo if possible. It reaches toward the lens to wipe the glass.
- The camera stays still. A steady, centered composition works best.
- Warm, soft lighting in the photo matches the warm blur of the steam.
Models this effect runs on
The same prompt on different models. Pick by look and by price.
- MiniMax H3$0.185 per run
5 s · 768p
- Seedance 2.0 Fast$0.484 per run
5 s · 720p
- Seedance 2.0$0.76 per run
5 s · 720p
- Wan 3.0$0.45 per run
5 s · 720p
- Vidu Q3 Turbo$0.33 per run
5 s · 720p
- Wan 2.6 Flash$0.113 per run
5 s · 720p
- Seedance 2.0 Mini$0.265 per run
5 s · 720p
How it works
- 01
Add one photo
Choose the photo to match the example: include both people for a shared gesture, and keep the face and any hands involved clearly visible. Upload your own photo or start with a Steamy Mirror sample.
- 02
Check the price
Steamy Mirror defaults to a 5-second clip for $0.185. Swap the model or change the length and resolution before you run — the price on the button updates to match, and that's the whole cost of the clip.
- 03
Generate
Your Steamy Mirror clip renders in the background, lands in History and is never published.
Do the move yourself
Steamy Mirror keeps one version of this shot. To give Steamy Mirror your own motion or ending, write it yourself in Image to Video, starting from the prompt above.
What it does to your photo
The camera is a mirror. Within the first second thick steam fogs the glass between the lens and the person until they are only a warm blur. By the second second they reach toward the lens and wipe one wide clear streak across the fogged glass with the flat of their hand, and through the clean strip their face appears sharp, looking straight into the lens with a knowing smile. Fresh fog creeps back in at the edges of the streak while the camera holds still. The person stays exactly as they appear in your photo throughout.
In our test runs on MiniMax H3, a man with wet slicked hair and a thick moustache in a misty, warm-lit room was blurred behind fogged, droplet-covered glass by second two, raised his palm and wiped it clear by second three, and appeared sharp through the cleared glass with a broad grin by second four. In a second run, a man with short dark hair stood behind a glass door that fogged over within the first second, then swept his palm across the glass by second two so his face came through sharp, ending with a knowing smile into the lens as fog crept back in at the edges.
Head-and-shoulders portraits with the face toward the camera work best, since the wiped streak reveals the face. A person with a hand able to reach forward helps the motion read. Photos where the face is turned far away or hidden make the reveal weak. Very dark photos hide the fog. The effect leaves the person exactly as they are in your photo and only adds the motion, light and atmosphere.
It runs on MiniMax H3 for five seconds by default, and H3 itself goes up to 15 seconds and 2K if you want more. Studio also lets you move the same prompt to another model before you generate, such as Seedance 2.0 or Wan 3.0, and the price on the Generate button updates with every change. Our samples were all made on MiniMax H3, so other models will read the prompt in their own way.
Good for moody, atmospheric portraits and short clips built around a reveal.
Open the prompt in the generator and change "one wide clear streak" to "a small clear circle" for a smaller opening, or "a knowing smile" to "a calm, steady look" for a quieter finish.
