Skip to main content

HeyGen Video 1.0 Prompting Guide

Prompt HeyGen Video 1.0
for scenes that speak

HeyGen Video 1.0 is HeyGen's first general video model. Every clip comes with sound, including spoken lines you write in quotes, ambience and effects. Here is how to write a prompt that gets all three.

Overview

Describe a scene and what the people in it say. HeyGen Video 1.0 films the scene and adds the voices, the room sound and the effects in the same clip. You do not need a separate step for audio.

Sound in every clip

Spoken dialogue, ambience and sound effects come with each clip. Describe the sounds you want in the prompt.

Lines you write

Put the words in quotes and say who speaks and how. They are spoken as written.

5 to 30 seconds

One clip can be a quick beat or a full scene. Pick the length that fits your script.

480p to 2K, six shapes

Choose 480p, 768p, 1080p or 2K. Shapes include 21:9, 16:9, 4:3, 1:1, 3:4 and 9:16. Start from text or one image.

Prompt Framework

Write the picture first, then the voice, then the sounds. One paragraph is enough.

1. Subject

Who or what is in the frame. Give a short look: age, hair, clothing.

2. Action

What they do, in one or two simple moves.

3. Camera

The shot and any move, such as a slow push in or a low tracking shot.

4. Setting

Place, time of day and light, so the scene has a clear mood.

5. The spoken line

The exact words in quotes, with who says them and how: calm, excited, low, cheerful.

6. The sounds you want

Ambience and effects you want to hear. Say "no music" if you want none.

Prompt template

Fill in each part. Keep the spoken line short enough for the clip length.

[Subject: who, age, look, clothing] [action] in [setting, time of day, light]. [Camera: shot and move.] [Speaker] says in a [tone] voice: "[the exact line]" Audio: [ambience], [effects], [no music if you want none].

Sample Prompts

Six prompts to copy and change. Each one shows the settings it was written for.

Talking-to-camera product pitch

8 s · 768p · 9:16

Vertical selfie video. A strikingly beautiful woman in her late twenties with glossy dark hair and a warm smile holds up a sleek matte-black water bottle in a bright, modern kitchen, morning sun on the counter behind her. She looks straight into the camera and says in a friendly, confident voice: "One bottle, all day cold. This is the only one I carry now." She taps the bottle with her nail and it gives a soft metallic ping. Audio: her clear close voice, the ping, quiet kitchen room tone.

Put the spoken line in quotes, describe how it is said, and name the sounds you want around it. Two short sentences fit an 8 second clip.

Two-person dialogue scene

8 s · 768p · 16:9

Medium two-shot in a dim jazz bar at midnight, amber light, rain on the window behind them. A handsome man in his thirties with short dark hair and a charcoal suit sits across a small table from a gorgeous woman with auburn curls and an emerald dress. They take turns, one voice at a time. He says in a low, warm voice: "You came." She smiles, lifts her glass, and answers softly: "I always do." Audio: a muted piano, rain on glass, glasses clinking, their two voices.

Place each speaker by what they wear and where they sit, give each line its own beat, and say they take turns so the voices stay apart.

Cinematic action with sound design

6 s · 768p · 16:9

A ruggedly handsome stunt driver in a black leather jacket slides a muscle car sideways around a wet hairpin turn on a mountain road at dusk. Low tracking shot beside the rear wheel, then a quick cut to a close-up of his focused face lit by the dashboard. Spray flies off the tires, headlights sweep the pines. Audio: a deep engine roar, tires screaming on wet asphalt, gravel pinging the underside, a sharp gear change, no music.

Write the sound like a second shot list: engine, tires, gravel, gear change. Naming each effect gives the clip a mix you can hear.

Street food ASMR

5 s · 768p · 1:1

Extreme close-up, shallow depth of field, a night market stall glowing with string lights. A charming, handsome vendor in a white apron drops a skewer of marinated chicken onto a hot grill and flips it with tongs. Fat sizzles, smoke curls up through the lights, he brushes on glossy red sauce and a drop falls and hisses on the coals. Audio: loud close sizzling, tongs clicking, a brush dabbing sauce, soft crowd murmur far behind.

Stay close and let one action carry the clip. Ask for loud, close sounds and a quiet background so the details stand out.

Ultra-wide landscape with ambience

8 s · 768p · 21:9

Ultra-wide cinematic shot of a misty alpine lake at sunrise, pine ridges fading into fog, glassy water. A strikingly beautiful woman in a red wool coat stands on a wooden dock at the right of frame, looking out over the water, her hair moving in the breeze. The camera drifts forward very slowly. Audio: gentle lapping water, a distant loon call, wind in the pines, no music, no dialogue.

Pick 21:9 for scenery and place your subject on one side of the frame. For ambience, list three or four layers and say there is no music or speech.

Stylized animation

6 s · 768p · 16:9

Colorful 3D animated style, soft rounded shapes, bright pastel palette. A beautiful young inventor with a silver bob and big round goggles pushes up her sleeves in a cluttered treehouse workshop and cranks a tiny brass robot to life. The robot blinks, hops onto the table, and squeaks. She grins and says in a cheerful, playful voice: "Welcome to the team, Bolt!" Audio: whirring gears, happy squeaks, her voice, a light wooden creak.

Name the style first and keep it consistent through the prompt. Cartoon sounds like squeaks and whirs suit the look.

Image to Video

Add one image as the start frame and write what happens next. The clip takes the shape of your image, so the aspect ratio setting does not apply. You choose the resolution and the length.

Describe the motion, not the picture

The image already shows the scene. Use the prompt for movement, camera and sound.

Give your subject a line

A portrait can speak. Quote the words and describe the voice.

Keep it to one action

A single clear move holds up better than a list of events.

Match the sound to the image

Name the ambience that fits the place in your picture.

Image-to-video template

Use with one start frame.

The camera pushes in slowly as [subject] [one action]. [Subject] says in a [tone] voice: "[short line]" Audio: [ambience from the picture], [one effect].

Pro Tips

Fit the line to the length

Two short sentences suit 8 seconds. For a longer script, raise the duration instead of adding words.

Say who speaks

With two people, give each line its own beat and say they take turns.

Try 480p first

Check the idea and the voice at 480p, then render the keeper at 768p or higher.

Pick the shape for the place it will be seen

9:16 for phones, 16:9 for video, 1:1 for feeds, 21:9 for scenery.

Control the music

Add "no music" when you want the room sound and the voices on their own.

Use a seed to compare

Keep the seed and change one word to see exactly what the word does.

Parameters

SettingValuesNotes
PromptUp to 5,000 charactersPicture, spoken lines and sounds.
Duration5-30 sWhole seconds.
Resolution480p, 768p, 1080p, 2KIterate low, finish high.
Aspect ratio21:9, 16:9, 4:3, 1:1, 3:4, 9:16At 1080p and 2K, 16:9 or 9:16. Not used with a start frame.
Start frameOne image (optional)The clip follows the image's shape.
SeedWhole number (optional)Repeat a result.
SoundAlways onDescribe it in the prompt.

FAQ

Any whole number of seconds from 5 to 30, in one clip. Longer clips have room for a longer line of speech or a slower scene.