Turn a single image and audio track into a talking, singing, or animated character video. Use optional prompts to guide facial expressions, body movements, and camera shots.
Upload a character image and audio track. Add optional instructions to guide the performance.
Watch OmniHuman 2.0 video examples and compare AI lip sync, facial expressions, body motion, singing performance and character consistency.
Watch the mouth shapes on fast phrases, natural pauses and emphasized words to see whether every sound stays aligned.
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.
Check whether the bottle keeps its shape and position as the hand lifts and lowers it, without warping, slipping or flickering.
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.
Watch how sustained notes, breaths and lyric timing remain synchronized while expression and body rhythm build with the music.
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.
Focus on how pointing, open-hand gestures, head movement and posture follow the meaning without looking repetitive or disconnected.
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.
Look for subtle eye focus, eyebrow response, blinking and changes around the mouth that match the speaker’s tone and intent.
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.
Observe whether the soft-clay surfaces, friendly face and character identity stay consistent through lip sync, leaf movement and articulated hand gestures.
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.
OmniHuman 2.0 is an audio-driven AI avatar video generation model from ByteDance and an upgraded version of OmniHuman 1.5, designed for highly dynamic character motion.
It turns a single reference image and an audio track into videos of characters speaking, singing, and moving. Optional text prompts can guide character actions, facial expressions, camera movement, and scene events.
Try OmniHuman 2.0 Online
Generated videoExplore how OmniHuman 2.0 uses audio and text prompts to create expressive character performances, guide movements, and control scenes.
Use speech or singing audio to drive mouth movements and character performances. Turn a reference image and audio track into an animated video.
Generate character actions based on the meaning of spoken audio. Characters can show objects, respond to dialogue, or react to others without an additional text prompt.
Use text prompts to guide character movements, camera angles, and background events. Describe multiple actions in sequence to direct how a scene unfolds.
Generate facial expressions and physical reactions that reflect emotions in speech and dialogue. Guide how characters respond as the tone of a scene changes.
Animate multiple people within the same scene. Use prompts to specify who should speak while other characters move and react naturally in the background.
Animate a variety of subjects, including people, cats, dogs, and stylized characters. Create audio-driven performances from different types of reference images.
Create longer character performances from audio tracks up to 5 minutes. The generated video duration follows the uploaded audio length.
Automatically match male and female voices in the audio with the corresponding characters in the reference image for more natural multi-person dialogue.
Create an AI character video in four simple steps. Start with an image and audio, add optional instructions, and generate your video online.
Upload a clear image of the person, animal, or stylized character you want to animate. Choose an image that meets the supported file requirements and clearly shows the subject.
Upload a speech recording, dialogue clip, or singing track. The audio provides the voice or performance for your generated video. Make sure the file meets the supported format, size, and duration limits.
Describe the actions, expressions, or camera movement you want to see. Use clear, sequential instructions to guide the performance. You can leave the prompt blank and let the audio guide the character's performance.
Review your uploaded files, confirm the input duration and displayed credit cost, and start generation. When processing is complete, preview your video and download the result if it meets your needs.
Explore practical ways to turn reference images and audio into videos for social media, education, storytelling, marketing, and creative projects.
Turn an existing voice recording and character image into short-form video content for social platforms. Add a concise prompt to shape the presentation, then use the generated clip as a starting point for creator updates, personal stories, or voice-led posts.
Pair a reference image with lesson narration to create a presenter-led explainer. Use a prompt to guide the presentation and visual actions, giving educators and content teams another way to turn prepared scripts into video lessons without filming every segment.
Combine character images with dialogue audio to develop scenes for original stories, fictional conversations, or serialized content. Use prompts to guide the intended actions, then review the generated clip as a draft for a larger character-led narrative.
Explore early video ad ideas by pairing a suitable character image with campaign narration or dialogue. Generate a draft to evaluate the creative direction before investing in a full shoot, detailed editing, or final campaign production.
Turn a prepared script into presenter-style content using a reference image and recorded narration. Add optional instructions for gestures or framing, then use the generated video as a draft for announcements, internal explainers, or educational presentations.
Pair a character image with a song or vocal track to create music-led clips for music promotion, creative concepts, or entertainment posts. Use the generated result to explore a visual direction for a track before developing a finished music video.
See how different prompts affect videos generated from the same image and audio. Compare the results, explore motion and emotion controls, and try the prompts yourself.
Compare how one change in the prompt affects the generated video.
Audio-driven baseline
No prompt used.
The audio produced natural speech and light hand gestures, but the blue card remained on the desk throughout the video.
Step-by-step movement control
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.
The character picked up the blue card and displayed it beside her face. She followed the main action, but continued holding the card at the end instead of returning it to the desk.
Movement and expression control
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.
The character showed a clearer emotional progression, picked up and displayed the blue card, then returned it to the desk before finishing with a confident smile.
Prepare a clear reference image and clean audio, write focused motion prompts, and test complex actions with shorter clips before creating longer OmniHuman 2.0 videos.
Create a VideoUse a sharp, well-lit image with visible facial details. Avoid heavy blur, extreme viewing angles, or very small faces. A clear reference gives the model better visual information to work with.
Use a clear voice or singing track with low background noise and minimal overlap from other speakers. Clean audio gives the model stronger timing, rhythm, and emotion cues for lip sync and character movement.
Choose an image that makes the character's face easy to see. Very distant shots may reduce lip-sync reliability. If the mouth does not move as expected, try a closer reference image or simplify the camera instructions.
Write prompts as a short series of events. Explain what the character does first, what happens next, and how the scene ends. Avoid conflicting instructions or too many unrelated movements in one prompt.
Keep the main character visible throughout the scene when facial and visual consistency matter. Repeatedly moving out of frame and reappearing later may lead to changes in appearance. Use simpler camera paths when character consistency is important.
Test complex movements, camera changes, or multi-character scenes with shorter clips before attempting longer generations. Short tests make it easier to evaluate motion, framing, and character consistency before spending credits on a longer generation.
Choose the credit pack that fits your workflow, pay once and create whenever inspiration arrives.
A simple starting pack for testing OmniHuman and creating short character videos.
More credits for creators producing social, educational, and promotional videos.
The highest-volume credit pack for studios, teams, and longer production workflows.
Credits work with OmniHuman 1.5 and 2.0. Both use 1 credit per second of generated video.
Prices exclude VAT and local taxes, calculated at checkout.
Purchased credits do not expire and are added to your account after payment.
Find answers about using OmniHuman 2.0 online, free credits, supported files, prompts, video length, troubleshooting, commercial use and privacy.
Try OmniHuman 2.0 →Upload a reference image and an audio track, then add an optional text prompt to guide character actions or camera movement. Review your settings and the displayed credit cost, then start generating your video.
Yes. New users receive free credits when they register and can use them to try video generation. If you want to create more videos after using your available credits, you can purchase a credit pack.
You can upload JPG, PNG, WebP, BMP, TIFF, GIF, HEIC, or HEIF images under 30 MB. Images must be 300–6,000 pixels wide and tall, with an aspect ratio between 0.4 and 2.5. Audio files must be in WAV or MP3 format, under 15 MB, and between 4 and 300 seconds long.
No. A text prompt is optional. Upload a reference image and an audio track to generate a video. You can add a prompt to guide character movements, expressions, or camera direction when you need more control.
OmniHuman 2.0 can generate videos up to 5 minutes long. The final video duration depends on the input audio and the generation settings available on this platform.
A distant face, unclear audio, or an extreme viewing angle may affect mouth movement. Try uploading a sharper image with a larger, clearly visible face. Use a clean audio recording and simplify the camera instructions. If the issue continues, test a shorter clip with fewer actions.
The credit cost is displayed before generation begins. If generation fails or is blocked, all reserved credits are returned to your account.
Yes, you can use videos generated with OmniHuman 2.0 for commercial projects, including marketing content, product promotions, and social media campaigns, subject to our Terms of Service. Make sure you have the necessary rights to all uploaded images, audio, and other reference materials. You are responsible for ensuring your final content complies with applicable laws and third-party rights.
Both versions support audio-driven character animation and optional text prompts. OmniHuman 2.0 accepts longer audio inputs and larger image files, with an emphasis on more flexible character and camera direction. Actual video quality depends on the source materials and scene complexity.
Your uploaded images, audio, prompts, and generated videos are used to provide the video generation service. Some files may be processed by third-party AI model and infrastructure providers to generate your videos. We do not sell your personal data, and you can delete your stored files and generation history through your account, subject to our Privacy Policy.
Turn one character image and voice track into an expressive AI video with synchronized lip movement, facial emotion and natural body motion.
Create a video ↗