AI production terms, without the fog.

The proportional relationship between image width and height, such as 9:16 for vertical video or 4:5 for an Instagram feed post.

Keeping the same recognizable identity, features, age cues, hair, and proportions across different generated scenes.

Prompt instructions that define what must remain unchanged between frames, shots, or generation passes.

A reference used to guide pose, depth, edges, composition, identity, or another structural property.

The approved still image used as the starting point for an image-to-video generation.

Generating motion from an existing still image while attempting to preserve its subjects and composition.

Unwanted changes to a person, product, object, or vehicle as the model generates variations or motion.

A reference and prompt structure designed to preserve the defining visual characteristics of a subject.

Editing a selected region of an image while preserving the rest of the composition.

A visually important frame that defines the start, end, or major state of a shot.

Instructions describing unwanted artifacts, changes, styles, or objects the generation should avoid.

Extending an image beyond its original boundaries while generating new surrounding content.

The ordered structure of a prompt: locks, subject, scene, composition, camera, lighting, style, and constraints.

Loss of the original visual requirements as a prompt becomes overloaded or later instructions conflict with earlier ones.

An uploaded image used to communicate identity, product design, style, composition, or another visual requirement.

A numeric starting value some models expose to make generation behavior more repeatable, though it does not guarantee identical output across every setting or model version.

An ordered plan describing each shot's angle, framing, action, movement, purpose, and duration.

An image used to guide visual language such as color, lighting, texture, or editorial treatment rather than subject identity.

Creating a still image from written instructions without requiring an image input.

Creating a moving sequence primarily from written instructions.

An advertisement designed to feel like creator-made, platform-native content rather than a traditional polished commercial.

A finished output or breakdown that demonstrates a production capability instead of only describing it.

A documented sequence of inputs, decisions, tools, generation steps, approvals, and outputs that can be repeated.