Skip to content

Prompting guide

How to prompt LTX Video

How to prompt LTX Video on Mage: a quick start, a fill-in template, and expert tips for the 120-word rule, start frames, camera moves, and fast drafts.

Morning coffee
Made with LTX VideoMorning coffee

Quick start

Your first great LTX Video prompt

Everything a beginner needs in one minute: what to know first, the rules that matter most, a template to fill in, and prompts to copy.

What to know first

LTX Video makes short, silent clips fast, for trying ideas cheaply. One thing matters more than anything else: on Mage, a prompt under 120 words is rewritten by a small built-in writer before the model sees it. Write 120 to 200 words in one paragraph that starts with the action, and the model gets your words instead of someone else's. Each clip holds one continuous action and one camera move.

The rules that matter most

  1. Write at least 120 words

    Shorter prompts are rewritten automatically and lose your details. Reach 120 words with how things look, the setting, and the light, never with repeated adjectives.

  2. Start with the action

    The first sentence is the main action itself: "A small penguin takes a red book from the shelf…". Not the setting, and not a style line.

  3. Be literal

    Spell out colors, textures, materials, and light ("dark, jagged rocks covered in patches of green moss") instead of mood words and film jargon.

  4. One action, one camera move

    A clip lasts a few seconds: no cuts, no second scene, no story told in beats.

  5. Draft on Distilled, finish on Dev

    LTX Video Distilled is the fast default. Once a prompt works, run the same words on LTX Video Dev for the keeper.

Fill in this template

Replace each part in square brackets with your own idea, and leave out a part you do not need.

Template
[The main action, in one sentence]. [The movements and gestures that follow]. [What the subject looks like]. [The setting and background]. [The camera's angle and movement]. [The light and the colors]. [A change or event at the end].
The template, filled in
A barista pours steamed milk into a cup of espresso, drawing a leaf pattern on the surface. She tilts the cup toward the pitcher, wiggles the spout gently from side to side, then lifts it in one smooth line through the middle of the pattern. She is in her twenties, with short bleached hair, silver rings on both hands and a black canvas apron over a white T-shirt. The cup is wide and pale blue, with a speckled glaze and a matching saucer. Behind her, a chrome espresso machine and shelves of glass jars full of coffee beans blur softly. The camera stays still at counter height in a close-up on her hands and the cup. Warm morning light comes through a window on the left, with soft shadows and a cream and brown palette. A thin curl of steam rises as she sets the pitcher down.
  • [The main action, in one sentence] Start with the action itself: a barista pours steamed milk into a cup of espresso, drawing a leaf pattern.
  • [The movements and gestures that follow] How the action unfolds, step by step: she tilts the cup, wiggles the pitcher, then lifts it through the middle.
  • [What the subject looks like] People and objects in literal detail: age, hair, clothing, colors, materials.
  • [The setting and background] What surrounds the subject, near and far: a chrome espresso machine, shelves of glass jars.
  • [The camera's angle and movement] One camera behavior in plain words: the camera stays still at counter height; the camera slowly pushes in.
  • [The light and the colors] Where the light comes from and the palette: warm morning light from a window on the left, cream and brown tones.
  • [A change or event at the end] A small final beat, if any: a curl of steam rises as she sets the pitcher down.

Prompts to start from

  • A waterfall, from text

    White water pours over the edge of a mossy cliff and plunges into a deep green pool below, throwing up a fine mist. The falling water splits into three ribbons halfway down, then rejoins in a churning cloud of foam that spreads in rings across the pool. The cliff is dark gray basalt, wet and glistening, with thick cushions of bright green moss and small ferns growing from every crack. Smooth round boulders line the edge of the pool, and tall tree ferns lean in from both banks. The scene is dominated by the contrast between the bright white water and the dark, wet rock. The camera stays still at the water's edge, looking up at the falls. Soft overcast light, deep greens and grays, with a faint rainbow in the mist near the base.

    Why it works: It opens with the action, describes every surface literally, keeps the camera still, and runs past 120 words, so your words reach the model as written.

  • A start frame brought to life

    A gray tabby cat on a sunny windowsill stretches its front legs, yawns widely, then curls back into a ball and closes its eyes. Its tail flicks once and settles over its nose as its sides rise and fall slowly. The cat has striped silver-gray fur, white paws, a pink nose and pale green eyes. The windowsill is painted white wood, with a small terracotta pot of basil at one end and a folded blue blanket under the cat. Through the window behind it, rooftops and a strip of blue sky are softly out of focus. The camera stays still at the height of the windowsill in a medium close-up. Warm afternoon sunlight falls across the cat from the right, with soft shadows and a gentle gold and gray palette.

    Why it works: Upload the photo as the First Frame. The prompt still opens with the action, then describes what the photo shows in the same literal words, so picture and prompt agree.

  • A hot-air balloon, from text

    A red and yellow striped hot-air balloon rises slowly from a misty valley at sunrise, its burner flaring orange as it lifts. The basket sways gently as the balloon climbs past the treetops, and two passengers in dark jackets lean over the edge and point toward the hills. The balloon is huge and round, with bold vertical stripes and a wicker basket hung on thick ropes. Below it, a patchwork of green fields and hedgerows stretches away, and a narrow river glints between rows of poplar trees. The camera follows from a distance at the same height, drifting slowly to the left as the balloon rises. The light is low and golden, with long shadows across the fields, pale blue mist in the hollows and a peach-colored sky.

    Why it works: One continuous action, one slow camera move, and literal detail on the balloon, the land, and the light. Draft it on LTX Video Distilled, then run the same words on LTX Video Dev.

Advanced

Advanced LTX Video prompting

Techniques for more control once the basics work.

Distilled or Dev

LTX Video Distilled is the default: fast, built for drafts and iteration. LTX Video Dev is slower and renders a higher-quality clip, so use it for the keeper once the prompt works. Both read the same prompt.

LTX Video is the model for trying ideas. When a shot works here, take it to a larger video model for the final, such as Kiwi, Lemon, or Cherry.

  • The Negative Prompt has little or no effect on Distilled; it counts on Dev.

Why 120 words matters

Mage runs LTX Video with its built-in prompt writer always on. Any prompt under 120 words is rewritten by a small language model before the video model reads it, and with a start frame that model first describes the image in its own words. A prompt of 120 words or more goes through exactly as you wrote it.

So a short prompt is not a simpler prompt; it is someone else's prompt. Writing 120 to 200 words in the order below keeps every detail yours.

  • Reach 120 words with real information: the subject's colors and materials, what is behind it, where the light comes from. Never pad with repeated adjectives.
  • Count before you generate: 120 words is roughly seven or eight full sentences.
  • Keep it under about 200 words: the essentials of each part, not every detail you can think of.

Structure: seven parts in one paragraph

Write one flowing paragraph, in time order, literal and precise, in English. Think like a cinematographer describing a shot list. The order that works best is: the main action in one sentence that starts with the action itself; the movements and gestures; the precise look of the people and objects; the background and setting; the camera angle and movement; the light and colors; and any change or sudden event.

The more elaborate the description, the better. Spell out colors, textures, materials, and light the way you would describe a photo to someone who cannot see it: "The rocks are dark and jagged, covered in patches of green moss."

  • For an establishing shot, follow the action sentence with "The scene is dominated by…" and the strongest visual contrast in the frame.
  • "The video depicts…" works as a second sentence for a wide view of a place.
  • Prefer plain description to jargon: "the waves crash against the rocks, sending white foam spraying into the air" beats "dramatic, epic surf".

Camera and motion

A clip holds one continuous action and one camera behavior, named in plain words. Describe motion the way it would look: the path, the speed, what moves and what stays still.

  • Everyday motion holds up best: steam, water, hands at work, leaves in the wind, a vehicle on a road.
  • Two actions in a row are fine when one flows into the other ("takes the book, opens it, starts reading"); a cut to another place is not.
WriteWhat it does
the camera stays still at eye levelA locked-off shot: only the subject moves.
the camera slowly pushes inThe camera moves gently closer to the subject.
the camera tracks alongside from the leftThe camera travels sideways with a moving subject.
the camera follows from high above and behindA chase view from above, good for vehicles and roads.
from a medium shot to a close-upThe framing tightens over the clip.

Starting from an image

Upload a still as the First Frame and the clip opens on it exactly. The prompt still opens with the action; then describe what the frame shows in the same literal terms (the colors, the clothing, the setting), so the words and the picture agree.

The clip takes the shape of the image, and the aspect ratio setting does not apply. Crop the image to the shape you want before you upload it.

  • For a character across several clips, start each one from an approved still of that character and repeat the same appearance sentence, word for word.
  • LTX Video takes no Characters, References, or @handles, and no last frame. An @ followed by letters stops the generation before anything is charged.

The negative prompt

The Negative Prompt setting starts filled in with "worst quality, inconsistent motion, blurry, jittery, distorted". Keep it. On LTX Video Dev, add anything specific to the shot; on Distilled, the negative prompt has little or no effect, so fix a draft with a clearer description instead.

Sound and text

LTX Video clips are silent. Leave sound and dialogue out of the prompt, and add audio afterwards, for example music or effects made with Seed Audio.

Text in the picture is not a strength. Add titles and captions in an editor.

Common mistakes

Avoid this, do this instead

  • Avoid: A one-line prompt.

    a penguin reading a book in a library

    Instead: 120 to 200 words in the seven-part order, so the built-in writer stays out of it.

    A small penguin takes a red book from a low wooden shelf, opens it with both flippers and starts reading. It tilts its head toward the pages, then turns one page slowly as its feet shuffle on the floor. The penguin is round and glossy, with a black back, a white belly, a short orange beak and bright black eyes. The book has a worn red cloth cover with faded gold lettering. Behind it, tall dark bookshelves full of old leather-bound books stretch back into a quiet library with a green-shaded lamp on a desk. The camera stays at the penguin's eye level and slowly pushes in from a medium shot to a close-up. Warm lamplight falls from the left, with soft shadows and a calm amber and green palette.
  • Avoid: Opening with the setting or a style line.

    Cinematic, moody, a misty forest at dawn where a deer…

    Instead: The main action as the first sentence.

    A young deer steps out of the trees and lowers its head to drink…
  • Avoid: Film jargon and mood adjectives.

    Instead: Literal descriptions of motion, colors, textures, and light.

  • Avoid: A story with several beats, or a cut.

    Instead: One continuous action.

  • Avoid: Expecting the negative prompt to fix a Distilled draft.

    Instead: A clearer positive description, or LTX Video Dev.

  • Avoid: More than 200 words.

    Instead: The essentials of each of the seven parts.

  • Avoid: Dialogue, music, or sound effects in the prompt.

    Instead: Nothing: the clip is silent. Add audio afterwards.

FAQ

LTX Video prompting questions

How long should an LTX Video prompt be?
120 to 200 words. Under 120, a built-in writer rewrites the prompt before the model reads it, so your details can be lost. Over 200, trim to the essentials of each part.
Why doesn't LTX Video follow my short prompt?
Because a prompt under 120 words is rewritten by a small built-in writer before the model sees it, and your details can get lost. Write 120 words or more and the model reads exactly what you wrote.
Does LTX Video make sound?
No. The clips are silent. Add music, voices, or effects afterwards, for example with Seed Audio.
Does LTX Video support negative prompts?
Yes, and the Negative Prompt setting comes with a good default. It matters on LTX Video Dev; on Distilled it has little or no effect.
How do I keep a character consistent in LTX Video?
Start every clip from an approved still of the character as the First Frame, and repeat the same appearance sentence word for word in each prompt.
Can LTX Video use Characters or reference images?
No. It takes a prompt and, optionally, one start image. For Characters and references, use a model such as Cherry or Lemon.
What is the difference between LTX Video Distilled and Dev?
Distilled is the fast default for drafts; Dev is slower and renders the higher-quality keeper. The same prompt works on both.

Make your first video with LTX Video

New accounts get 300 free Gems, enough to try LTX Video Distilled. No install, no credit card.

Try LTX Video Distilled