OmniHuman Writing Guide
Learn how to write effective prompts for OmniHuman 1.5 and 2.0. Guide character movements, facial expressions, camera actions, and scene events with clear, step-by-step instructions.
This guide covers prompt-writing techniques for both OmniHuman 1.5 and OmniHuman 2.0. Explore practical examples for character actions, emotional performances, camera movement, and multi-character scenes.
Prompt support and generated results may vary between model versions. Use the examples as a starting point and adjust your instructions based on the model you choose.
Camera movement + Emotion of the speaking character + Speaking state (talking/crying/singing/...) + Specific actions + (Optional) Background events/Actions of other characters
Complete Example: 'The camera slowly moves from the side to a medium-front shot. A young woman is sitting by the window, calm, gently picking up the coffee cup and smiling as she talks to the camera. A boy beside her quietly watches her and occasionally turns to the camera and smiles. The lights in the background are gently flickering.'
Note: Writing 'Character xxx talks' or 'Character xxx sings' in the prompt helps improve the speaking performance.
We used the same reference image and audio in OmniHuman 2.0 and 1.5, starting without a prompt, then adding an action sequence and emotional direction. Compare the actual videos to see what each instruction changed and where the result fell short.
Original website generation · Input and output reviewed
No prompt used
The presenter uses light conversational gestures while the blue card remains on the desk throughout the video.
The presenter speaks with natural head and hand movement. Her hand moves toward the card, but she does not deliberately present it.
Original website generation · Input and output reviewed
The woman speaks to the camera. During the first sentence, she keeps both hands resting on the desk. During the second sentence, she picks up the blue card with her right hand and holds it beside her face while continuing to speak. During the final sentence, she places the card back on the desk and gives one small confirming nod as she finishes speaking.
She picks up and displays the card, but continues holding it at the end instead of returning it to the desk.
She picks up the card, holds it beside her face and returns it to the desk. Most of the requested sequence appears in the right order.
Original website generation · Input and output reviewed
The woman speaks to the camera. During the first sentence, she keeps both hands resting on the desk and speaks with a thoughtful, slightly uncertain expression, briefly drawing her eyebrows together. During the second sentence, her eyes brighten and her expression becomes more optimistic as she picks up the blue card with her right hand and holds it beside her face while continuing to speak. During the final sentence, she places the card back on the desk, changes into a calm and confident smile, and gives one small confirming nod as she finishes speaking.
Her expression moves clearly from uncertainty to optimism and confidence. She also displays and returns the card before the final smile.
She displays the card, returns it to the desk and finishes calmly. The requested emotional change is visible, but more subtle.
Use no prompt when natural speech and spontaneous gestures are enough. The model did not deliberately present the blue card without an instruction.
Connect actions to moments in the audio: first, during the next sentence, and finally. Both versions showed the card interaction; 1.5 completed more of the requested sequence in this example.
Layer expression changes onto the established sequence. Both versions followed the main action, while 2.0 showed the emotional progression more clearly in this example.
Original website test · OmniHuman 2.0 and 1.5 · Results describe these specific generations and are not a guarantee of identical output.
These examples were created with original images, audio and prompts used on this website. Each one shows a different way to connect spoken timing with visible performance.
Use a locked close-up with direct eye contact. Form each phoneme clearly and precisely, fully close the lips during the pause, and add a natural breath before the final phrase. Keep head motion minimal and avoid large gestures so lip synchronization remains easy to evaluate.
The generated mouth movement stays closely aligned through the pace change and deliberate pause. The quickest consonant transitions remain the most useful area to inspect at full playback speed.
On “left,” indicate one side with the left hand. On “right,” shift gaze and open the right hand toward the opposite side. On “together,” move both hands smoothly toward the center, then return to a relaxed posture. Keep shoulders, arms, head and gaze coordinated without abrupt or repetitive motion.
Gestures follow the spoken cues with coherent head and posture changes instead of repeating one motion. Fine hand and finger transitions may appear less stable during the fastest movement.
Begin with slightly uncertain eye focus and a gentle brow pinch. When the result appears, brighten the eyes and lift the eyebrows. After the pause, transition into a natural smile with a soft exhale, then settle into a warm, satisfied expression. Avoid exaggerated acting and large body movement.
Eye focus, blinking and eyebrow movement support the change in tone without becoming exaggerated. Quieter phrases produce subtler cheek and mouth motion than the more emphatic lines.
Perform the lyrics with accurate singing lip sync and controlled breathing. Begin still and reflective, place one hand near the heart on the second line, then open the hand outward as the emotion builds. Add a gentle body sway while preserving the stage and fixed camera.
Sustained vowels remain visually connected to the vocal line while expression and body rhythm build with the music. Fast lyric transitions and short breaths are less precise than the held notes.
On “lift,” raise the bottle smoothly with one hand. On “turn,” rotate it slightly. On “steady,” hold it completely still. On “set it down,” lower it slowly to the starting position. Preserve finger contact, bottle shape, scale, label area and occlusion throughout; avoid warping, slipping or flicker.
The bottle keeps a consistent overall shape and remains connected to the hand through the lift, turn and lowering sequence. Finger contact and object edges should still be checked frame by frame during the rotation.
During the thoughtful opening, tilt the head slightly and let the single leaf antenna lean gently. When mentioning two shapes, indicate each tabletop object with the corresponding hand. On “Ah,” blink naturally and brighten the simple smile. On “your turn,” make one small inviting hand gesture. Preserve the soft-clay surfaces, rounded head shape, facial placement and articulated joint structure throughout.
The character identity, rounded proportions and soft-clay appearance remain recognizable while the face, leaf and hands animate together. Small facial and joint details may soften when several motions overlap.
Explore eight prompt techniques with OmniHuman 1.5 examples. Each example keeps its reference image, audio, prompt and generated video together.
Use "first… then…" to choreograph multi-step actions.
Always include "talking" or "singing" if the character must speak.
"The man walks forward first, then stops and puts his hands on his hips while speaking.
Then he turns around to look at the explosion behind him, showing his back.
The clothes on his back have been blown to pieces."
"The camera follows. The man turns around, looks at the camera and walks forward, singing.
He first touches his collar with both hands, then spreads his hands and raises his head."
"The man quickly took the cigarette out of his mouth, then started speaking to the camera."
"The camera zoomed in. The woman spoke to the camera, and after finishing, turned around and ran backward."
Use the voices and pauses in the audio to make each speaking turn clear. Describe what the other person does while listening.
Use a locked eye-level medium two-shot and preserve both character identities throughout the scene. The woman on the left speaks first with natural mouth movement, eye contact toward the man and one small open-hand gesture. The man remains silent, listens attentively, gives a subtle nod and does not move his lips. After the brief pause, the man speaks while turning his attention slightly toward the woman. The woman stops speaking, keeps her mouth closed and gives a natural listening reaction. During the final line, the woman speaks again and gives a small confirming nod while the man returns to a listening expression. Keep both people visible at all times. Maintain stable faces, clothing, hands, table objects, lighting and background. Do not move the cups. Avoid camera movement, exaggerated gestures, overlapping speech, identity changes or additional people.
What to watch: Watch how the active speaker changes while the listening character maintains natural eye contact, facial reactions and stable body posture.
"Following the camera, the boy quickened his pace to catch up with the girl and then started talking to her.
The girl stopped, turned around impatiently, and stared at the boy."
"Following the camera, the man holding a torch on the left looked at the others while speaking.
Then he turned back to look into the camera."
Precisely control movement type, speed, and end position.
Use "rapidly / quickly" for stronger responsiveness.
"Pan the camera to the right in a circular motion.
The character speaks while facing forward.
Then quickly zoom in and focus on the character's front face.
The character puts down the binoculars and speaks."
"First, the man looked calm and spoke to the camera.
Then he suddenly became nervous and panicked, frowning tightly."
"First, the woman rolled her eyes disdainfully and spoke with contempt.
Then she supported her chin with her index finger.
Finally, she played with the box on the table with a proud expression."
"Distorted air ripples appear.
The father and son are sucked into the depths like dust and disappear instantly,
leaving only the remote control and phone on the empty sofa."
"The chick puts on sunglasses, holds two guns, and speaks evilly."
"The reporter reports to the camera.
The king behind collapses.
The camera focuses on the collapsed king.
The reporter quickly turns around and speaks while looking behind."
"The woman turns around and speaks while looking at the soldier behind.
The camera focuses on the soldier.
Then she turns back to speak to the camera."
"A gentle breeze blows.
The character's clothes and hair flutter slightly.
The leaves in the background sway."
"The orangutan maintains a selfie pose and talks to the camera.
The volcanic lava behind gradually turns blue."