How do I describe camera angles and shot types in a prompt?
To describe camera angles and shot types in an AI prompt, use standard cinematography terms: name the shot size (extreme close-up, medium shot, wide shot, establishing shot), the camera angle (eye level, low angle, bird's-eye, Dutch tilt), and the movement (static, pan, dolly push-in, crane rise, handheld). These terms map directly to patterns the model learned from real film and photography, so they reliably steer composition and framing without needing long descriptions.
Shot size controls how much of the scene you see
Shot size is the most immediate framing lever. An extreme close-up (ECU) fills the frame with a single detail — an eye, a product logo, a fingertip on a keyboard. A close-up (CU) shows the face from the neck up, good for emotion. A medium shot (MS) frames from the waist up, the standard conversation frame. A full shot includes the entire body. A wide shot (WS) shows the subject in context of the environment. An establishing shot (ELS) goes wider still to orient the viewer in a location. Naming the shot size in your prompt is often the fastest way to fix a composition that feels wrong.
Camera angle changes the power dynamic
The angle the camera takes relative to the subject is one of cinema's most powerful tools and one of the easiest to invoke in a prompt. Eye level is neutral and observational — the most common default. A low angle (camera below the subject, pointing up) makes the subject look powerful or imposing. A high angle (camera above, pointing down) shrinks the subject and can feel vulnerable or surveilled. A bird's-eye or top-down shot is purely graphic and abstract. A Dutch tilt (camera rotated so the horizon is diagonal) creates unease or dynamism. Just naming the angle shifts the entire mood of the output.
Camera movement adds cinematic energy
For video prompts, movement vocabulary is essential. A static (locked-off) shot is stable and observational. A pan moves the camera horizontally on its axis. A tilt moves it vertically. A dolly (or truck) physically moves the camera forward, backward, or sideways — a dolly push-in creates intimacy and focus. A crane or jib shot rises or falls on a vertical arc. A handheld shot mimics the weight and shake of a human operator, adding naturalism or urgency. Drones give sweeping aerial movement. Specifying one of these in a video prompt is usually the difference between a clip that drifts randomly and one that feels directed.
Combine angle, size, and movement for precise direction
The most precise prompts combine all three layers: "low-angle close-up, slow dolly push-in toward the subject" gives the model a complete cinematographic instruction rather than a vague mood. Think of it like briefing a camera operator — they need to know where to stand (angle), what lens to use (size), and how to move. You can practice this by describing any memorable shot from a film or ad you admire and testing whether the model can reproduce the visual feeling. Floniks AI Video and the Workflow Editor both accept this vocabulary, and the /learn hub on Floniks has a visual glossary if you want to expand your shot language.
Related questions
Build it on Floniks
Image, video, digital humans, and reusable workflows on one canvas. Sign up gets you starter credits — no card required.
Explore Floniks