Video: LTX, Veo, Sora, Kling

Basic

Step 1 of 5

LTX-2 — narrative shots with camera and audio

Best use cases: Story-driven clips where camera behavior, character blocking, and sound (dialogue, ambience, music) should stay coherent together. LTX documentation emphasizes a complete scene from start to finish—not a keyword list.

Core elements (from LTX blog and API guide):

Element	What to write
Shot	Genre and scale (wide / medium / close-up); film language early if it matters.
Scene	Light, palette, materials, atmosphere.
Action	One continuous sequence in present tense.
Characters	Age, wardrobe, hair, distinguishing details; emotion through posture and face, not only adjectives.
Camera	When it moves, how, and what appears after the move.
Audio	Ambience, SFX, speech in quotes; language or accent if it matters.

Patterns that work: One flowing paragraph, present tense, about 4–8 sentences for a single coherent scene; match detail level to shot scale (close-ups need more specificity than wides).

Avoid (per LTX): Relying on abstract emotion labels without visual cues; expecting crisp readable signage; overloading many characters or chaotic physics in one prompt—iterate instead.

Example to adapt:

“Wide cinematic shot at golden hour on a coastal highway; a vintage coupe speeds toward camera with dust trailing; handheld tracking as it passes; camera pans left to hold the car as it drifts through the turn; tyre and wind noise, no music.”