Text2Movie

The film language each AI video model's docs let you write

Directing Open-Sora 2.0: a motion score in the prompt

Open-Sora 2.0's docs cover 3 of the six points. Framing: shown in examples. Frames: first frame only. Motion amount: a motion setting. Nothing on camera moves, cuts and negative prompt. As of 2026-09-25.

Camera moves per model, this one markedFive ways occur, from prompt words to drawn trajectories.Camera moves per model, this one markedCamera movesAdobe Firefly — Camera moves: Commands or presetsAdobe FireflyGoogle Veo — Camera moves: Words in the promptGoogle VeoGrok Imagine — Camera moves: Words in the promptGrok ImagineHiggsfield — Camera moves: Commands or presetsHiggsfieldHunyuanVideo 1.5 — Camera moves: Words in the promptHunyuanVideo 1.5Kling AI — Camera moves: Words in the promptKling AILeonardo Motion 2.0 — Camera moves: Commands or presetsLeonardo Motion 2.0LTX — Camera moves: Words in the promptLTXLTX Studio — Camera moves: Commands or presetsLTX StudioLuma Dream Machine — Camera moves: Commands or presetsLuma Dream MachineMidjourney — Camera moves: None in the docsMidjourneyMiniMax Hailuo — Camera moves: Commands or presetsMiniMax HailuoMoonvalley Marey — Camera moves: Drawn or set from an imageMoonvalley MareyOpen-Sora 2.0 — Camera moves: None in the docsOpen-Sora 2.0Pika — Camera moves: Words in the promptPikaPixVerse — Camera moves: Commands or presetsPixVerseRunway — Camera moves: Words in the promptRunwaySceneMixer — Camera moves: Words in the promptSceneMixerSeedance — Camera moves: From a referenceSeedanceSkyReels — Camera moves: Words in the promptSkyReelsSora — Camera moves: Words in the promptSoraStep-Video-TI2V — Camera moves: Words in the promptStep-Video-TI2VVidu — Camera moves: None in the docsViduWan — Camera moves: Words in the promptWan
Fig. 1 Five ways occur, from prompt words to drawn trajectories.
How each model asks for a camera move, read 2026-09-25.
ModelCamera moves
Adobe FireflyCommands or presets
Google VeoWords in the prompt
Grok ImagineWords in the prompt
HiggsfieldCommands or presets
HunyuanVideo 1.5Words in the prompt
Kling AIWords in the prompt
Leonardo Motion 2.0Commands or presets
LTXWords in the prompt
LTX StudioCommands or presets
Luma Dream MachineCommands or presets
MidjourneyNone in the docs
MiniMax HailuoCommands or presets
Moonvalley MareyDrawn or set from an image
Open-Sora 2.0None in the docs
PikaWords in the prompt
PixVerseCommands or presets
RunwayWords in the prompt
SceneMixerWords in the prompt
SeedanceFrom a reference
SkyReelsWords in the prompt
SoraWords in the prompt
Step-Video-TI2VWords in the prompt
ViduNone in the docs
WanWords in the prompt

Inclusion rule. All 24 models on the register, each with the way its own docs ask for a camera move. Order. Alphabetical by model name.

Camera moves
None in the docs
Framing
Shown in examples
Cuts
None in the docs
Frames
First frame only
Negative prompt
None in the docs
Motion amount
A motion setting

1What the docs say

The Open-Sora repository documents a motion score, 4 by default, that is fed into the prompt, with a dynamic option that lets a scorer pick it. Image-to-video takes a prompt and a reference image, and a sample prompt describes a bird's-eye view. Named camera moves, cuts and negative prompts are not documented.

How a camera move is asked for: None in the docs

How a camera move is asked for, every model

Shot sizes and angles in the docs: Shown in examples

The camera angle provides a bird's eye view of the waterfall.

HPC-AI Tech, Open-Sora repository

Open-Sora 2.0, framing in detail · Shot sizes and angles in the docs, every model

Cuts inside one generation: None in the docs

  • Repository read; several shots inside one generation are not described
    Not described in the docs read (2026-09-25)HPC-AI Tech, Open-Sora repository / recorded 2026-09-25

Cuts inside one generation, every model

Start and end frames: First frame only

Given a prompt and a reference image, you can generate a video with the following command

HPC-AI Tech, Open-Sora repository

Open-Sora 2.0, frames in detail · Start and end frames, every model

A negative prompt: None in the docs

A negative prompt, every model

A setting for how much the picture moves: A motion setting

During inference, you can use the following command to generate videos with motion score (the default score is 4)

HPC-AI Tech, Open-Sora repository

Open-Sora 2.0, motion amount in detail · A setting for how much the picture moves, every model

2Camera words in its docs

bird's-eye view

3Neighbouring models

How the docs were read: the method.

4Docs read