Correction, 2026-10-01: an earlier version said Veo 3.1 and Kling v3 render dolly moves, parallax and pull-backs convincingly and that Veo approximates a dolly-zoom at standard settings. We have not measured those results; the examples are now prompt advice. It also put clip lengths inside the prompts: on 8frame, length is a node setting (Veo 4, 6 or 8 seconds; Kling v3 3 to 15 seconds). Its Kling example read as text-to-video; on 8frame Kling v3 animates a start image, which the example now says.
A dolly shot is a camera move where the entire camera physically travels toward or away from a subject on a track, wheeled platform, or slider, producing smooth, continuous motion that no pan or zoom can replicate.
The difference between a dolly and a zoom matters more than it sounds. A zoom changes the focal length; the background scales but the camera stays put. A dolly moves the whole camera through space; the parallax shift between foreground and background layers changes, giving the viewer a genuine sense of depth and presence. That quality is why directors reach for a dolly when they want the audience to feel like they're entering a scene rather than watching it through a window.
How a dolly shot works
A dolly move has two basic directions: dolly-in (camera travels toward the subject) and dolly-out (camera moves away, also called a pull-back). Both are smooth and motivated by the subject's emotional weight in the scene.
Dolly-in creates intimacy. The frame tightens, the subject grows, background elements drift apart. It's used to signal that something important is about to happen or to draw the viewer into a character's emotional state.
Dolly-out creates distance or revelation. The camera retreats, context expands around the subject, and the viewer understands their place in a larger world. Pull-backs often end scenes, or open them to reveal scale.
There's a third technique that combines the dolly with a simultaneous zoom in the opposite direction: the dolly-zoom, or Vertigo effect, named for Hitchcock's use of it in 1958. The camera dollies out while the lens zooms in (or vice versa), keeping the subject the same size in the frame while the background warps. The result is spatial disorientation used to signal psychological unease or sudden dread.
In traditional filmmaking, a dolly shot requires a physical track, a wheeled camera platform, and a dolly grip coordinating with the camera operator. In AI video generation, you describe the move in the prompt and the model attempts to synthesize it from what it learned in training.
When you use a dolly shot
Use a dolly-in when you need to signal emotional escalation without a cut. A character receiving difficult news, a product reveal where the final frame should feel earned, an opening title sequence that wants to feel cinematic rather than static.
Use a dolly-out (pull-back) when you want to establish scale, end a moment, or leave the audience with context after a tight emotional scene. A pull-back that reveals a crowd around a lone figure, or a product in the environment it belongs to, communicates more than a static wide shot.
Use a dolly-zoom when you want deliberate visual unease. It's a recognized cinematic shorthand that audiences read immediately, so reserve it for moments that justify the effect.
Examples in AI prompts
Veo 3.1 dolly-in. Name the move and give it layers to travel past: "slow dolly-in toward a woman reading at a cafe window, late afternoon light, table edge in the foreground, shallow depth of field." Set 8 seconds in the node (168 credits on Veo 3.1 Fast with sound). The parallax between the figure and the window is the tell that separates a real dolly from a zoom, so look for it in the take; if the frame only scales, try "the camera physically moves forward" in the prompt. See the Veo 3 prompt guide for how to structure motion language alongside lighting and lens descriptors.
Veo 3.1 dolly-zoom. For the Vertigo effect, state both directions: "dolly-out zoom-in, subject stays the same size and centered in frame, background stretches, psychological tension." This is a hard move to get from any generator, so plan for several takes and judge whether the background actually warps. A 4-second take on Veo 3.1 Fast is 84 credits. See Veo 3 prompts for cinematic shots for chaining it with other moves.
Kling v3 pull-back reveal. Kling v3 on 8frame animates a start image, so begin with a close still of the mug, then prompt: "slow pull-back from a coffee mug on a table, camera retreats to reveal a morning kitchen, natural light, cinematic color grade." Set 6 seconds in the node; Kling v3 Standard bills per second, so 6 seconds with sound is 103 credits (5 seconds is 86). Check that the kitchen stays consistent as it comes into view before you use it as an ad opener.
FAQ
Is there a camera-move setting for dolly shots on 8frame? On Veo, no: the move goes in the prompt. Duration, aspect ratio and sound are the node settings.
How long can a dolly clip be? Veo 3.1 runs 4, 6 or 8 seconds (8 when you start from an image); Kling v3 runs 3 to 15 seconds.
Related concepts
- Veo 3 prompt guide covers how to write motion prompts for Veo 3.1, including dolly moves, pans, and orbit shots.
- Veo 3 prompts for cinematic shots shows how to combine dolly moves with lighting, framing, and lens choices.
Sources
- Veo and Kling v3 durations, node settings and prices (calculateVeo31Cost, calculateKlingV3Cost): 8frame tool registry, read 2026-10-01
Ready to generate a dolly shot? Open the canvas on 8frame and run a Veo 3.1 prompt with a slow dolly-in.
Try it without signing up: the 8frame canvas opens on a real board. The canvas is free and unlimited; generation is paid from $19/month.