Text2Movie

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

Open-Sora 2.0: framing

On shot sizes and angles in the docs, Open-Sora 2.0 reads “Shown in examples”. Camera angle appears in the sample prompts, such as a bird's-eye view of a waterfall. As of 2026-09-25.

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

HPC-AI Tech, Open-Sora repository

1Beside the other models

6 other models read the same way: HunyuanVideo 1.5, Kling AI, LTX, SceneMixer, SkyReels and others.

Nothing past the docs is assumed about how the model behaves.

2About the field

Framing words are the cheapest direction there is: close-up or wide changes what a shot says without any setting. A glossary also tells a writer where the maker draws the line between sizes.

Framing, this model markedDark marks are points the docs do not cover or could not be read.Framing, this model markedReadingAdobe Firefly — Reading: Options to pick fromAdobe FireflyGoogle Veo — Reading: Defined in a glossaryGoogle VeoGrok Imagine — Reading: None in the docsGrok ImagineHiggsfield — Reading: None in the docsHiggsfieldHunyuanVideo 1.5 — Reading: Shown in examplesHunyuanVideo 1.5Kling AI — Reading: Shown in examplesKling AILeonardo Motion 2.0 — Reading: Options to pick fromLeonardo Motion 2.0LTX — Reading: Shown in examplesLTXLTX Studio — Reading: Options to pick fromLTX StudioLuma Dream Machine — Reading: None in the docsLuma Dream MachineMidjourney — Reading: None in the docsMidjourneyMiniMax Hailuo — Reading: None in the docsMiniMax HailuoMoonvalley Marey — Reading: None in the docsMoonvalley MareyOpen-Sora 2.0 — Reading: Shown in examplesOpen-Sora 2.0Pika — Reading: None in the docsPikaPixVerse — Reading: None in the docsPixVerseRunway — Reading: None in the docsRunwaySceneMixer — Reading: Shown in examplesSceneMixerSeedance — Reading: None in the docsSeedanceSkyReels — Reading: Shown in examplesSkyReelsSora — Reading: Shown in examplesSoraStep-Video-TI2V — Reading: None in the docsStep-Video-TI2VVidu — Reading: None in the docsViduWan — Reading: Defined in a glossaryWan
Fig. 1 Dark marks are points the docs do not cover or could not be read.
Shot sizes and angles in the docs, all 24 models, read 2026-09-25.
ModelReading
Adobe FireflyOptions to pick from
Google VeoDefined in a glossary
Grok ImagineNone in the docs
HiggsfieldNone in the docs
HunyuanVideo 1.5Shown in examples
Kling AIShown in examples
Leonardo Motion 2.0Options to pick from
LTXShown in examples
LTX StudioOptions to pick from
Luma Dream MachineNone in the docs
MidjourneyNone in the docs
MiniMax HailuoNone in the docs
Moonvalley MareyNone in the docs
Open-Sora 2.0Shown in examples
PikaNone in the docs
PixVerseNone in the docs
RunwayNone in the docs
SceneMixerShown in examples
SeedanceNone in the docs
SkyReelsShown in examples
SoraShown in examples
Step-Video-TI2VNone in the docs
ViduNone in the docs
WanDefined in a glossary

Inclusion rule. Every model on the register, whether or not it has a page of its own on this field. Order. Alphabetical by model name.

3Next on the reel