Skip to content

LongCat Video AI

Type a scene and let it run long: text to video, image to video, or continue a clip you already have.

The last moment of your shot, in good light

LongCat-Video model

One model. Three ways to make a long take.

A 5-second 16:9 still frame: wide view of a lone cyclist in a dark jacket on an empty coastal road at first light, grey-blue sea on one side and green hills on the other.
LongCat video AI long take of a cyclist on a coastal road

LongCat-Video

Best for: Long takes that keep the same subject and the same light from first frame to last.

Made by
Meituan
Weights
Open weights
Modes
Text, image
Resolution
720p

Pick it in the composer. The exact cost of a render is shown before you start it.

Start creating

Video continuation

One shot, continued four times.

Each segment picks up where the last one ended. Same road, same rider, light rising from blue night to soft gold.

00:0000:1500:3000:45
00:00A 5-second 16:9 still frame: a rising crane view from low to high, tilting down over wet black rocks at a shore, a white lighthouse standing in thick sea fog beyond.

Opening prompt: slow dolly out from a rocky shore to a lighthouse in fog

00:15A 5-second 16:9 still frame: slow pan right across a quiet coastal road as it leaves the sea and bends into a sleeping town of low stone houses with shuttered windows, empty street, no people.

Continue: the road leaves the coast and runs into a sleeping town

00:30A 5-second 16:9 still frame: a lone cyclist in a dark jacket rides into frame on a hill road as the sun clears the ridge behind, low golden-hour sun from behind and to one side, a glowing rim along th

Continue: a cyclist enters frame and rides on as the sun clears the hills

00:45A 5-second 16:9 still frame: the road opens onto a wide sweep of coast, an empty asphalt curve above a calm silver sea, distant headlands in haze.

Continue: the road opens onto a wide salt flat in full morning light

Opening prompt: a vintage 4x4 drives along a savanna ridge at dusk, camera tracking beside it, one lone acacia ahead

LongCat video prompts

Eight long takes to start from.

Pick a mode, copy a prompt or start creating. Each one is written to hold one subject and one light from the first frame to the last.

  • Ruins at dawn

    Slow pan across weathered stone columns and a broken arch on a grassy plain, cloud shadows crossing the ground, no cuts

    text to video · travel

    Make thisCopyAsk AI

  • Forest encounter

    A hiker in a blue jacket holds still as a T-rex steps out of redwood mist, camera slowly pushing in, one continuous shot

    text to video · story

    Make thisCopyAsk AI

  • Clay sneaker reveal

    Start from my sneaker photo: it turns on a turntable, then the camera pulls back to show the whole set

    image to video · product

    Make thisCopyAsk AI

  • Camera on the desk

    Start from my photo: a slow quarter orbit around a vintage film camera and a stack of prints on a walnut desk

    image to video · product

    Make thisCopyAsk AI

  • Dead forest walk

    Continue my clip: keep moving through the blue mist between the dead trees without cutting, same light, same pace

    continue a clip · story

    Make thisCopyAsk AI
  • A 5-second 16:9 still frame: a single pale translucent jellyfish with trailing tentacles drifts through pure black water, shallow depth of field, the bell sharp and tiny particles soft behind it.

    Deep water drift

    Continue my clip: the pale jellyfish keeps drifting through black water as the camera follows it slowly left

    continue a clip · nature

    Make thisCopyAsk AI
  • A 5-second 16:9 still frame: locked-off static shot of a dark tornado funnel touching down on a flat green plain under a black storm sky, lightning flickering in the clouds.

    Storm touches down

    Locked-off shot of a dark tornado touching down on a green plain under a black storm, lightning flickering in the clouds

    text to video · nature

    Make thisCopyAsk AI

  • Wheat field funnel

    Continue my clip: the tornado moves across the golden wheat field while the storm-teal sky darkens, camera stays still

    continue a clip · travel

    Make thisCopyAsk AI

LongCat video AI questions

LongCat-Video is an open-weights video model from Meituan: it makes clips from text, from an image, or by continuing a video you already have. It is built for long takes that keep the same subject and light.

LongCat-Video is built for long, coherent takes at 720p. How far a shot runs depends on the scene and your settings, and the exact cost of a render is shown before you start it.

Yes. Video continuation picks up from the last moment of your clip and carries the subject and light forward. Frameo makes one video per run, so you extend a shot one step at a time.

Continue a shot

There is no setup and no GPU to find. LongCat-Video sits on one canvas beside every frontier model, and credits pay for generation, not seats. You retain ownership of what you make. The Terms of Use set out the full position.

Your next take

Start a take that keeps going.

Describe one shot, or the clip you want to continue. Frameo opens with your words already in the box.

The last moment of your shot, in good light

Enter to send · Shift + Enter for a new line