Turn a still photo into a talking video with your own audio. Create lip-synced speech, facial expressions, and natural movement with OmniHuman 2.0. Add an optional prompt to guide the performance.
Upload a clear photo and a recording of the words you want it to say. Add optional motion instructions, review the credit cost, and generate your talking video.
JPG, PNG, WebP, BMP, TIFF, GIF, HEIC or HEIF; under 30 MB. Each side must be 300–6,000 px, with a 0.4–2.5 aspect ratio.
MP3 or WAV, 4–300 seconds, under 15 MB. The uploaded recording supplies the spoken words; typing a script does not generate speech.
The cost is 1 credit per second of uploaded audio, rounded up. See Pricing for credit packs.
You only need a photo and a speech recording to create a talking video. Follow these three steps to upload your files, customize the movement, and preview the result.
Select a portrait where the face and mouth are easy to see. Use a front-facing or slightly angled image with good lighting. If you plan to add hand movements, choose a photo that shows the hands.
Upload an MP3 or WAV recording containing the speech you want your photo to perform. Listen to the audio before uploading to check the words, timing, and sound quality. You can also add an optional prompt for gestures or expressions.
Review the estimated credit cost, then start generating. When your talking photo is ready, watch the video to check the lip sync and facial movements before downloading it.
See how one photo and the same audio recording produce different results with and without motion prompts. Compare the visible gestures, expressions, and limitations before creating your own talking video.
The presenter uses light conversational gestures while the blue card remains on the desk throughout the video.
Motion prompt: No prompt
What to notice: In this sample, the photo speaks with light gestures without any action instructions.
She picks up and displays the card, but continues holding it at the end instead of returning it to the desk.
Motion prompt: 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.
What to notice: A requested multi-step action needs review: she lifts the card but does not put it back.
Her expression moves clearly from uncertainty to optimism and confidence. She also displays and returns the card before the final smile.
Motion prompt: 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.
What to notice: In this sample, the change in expression and the return of the card are both visible.
These notes describe the videos shown here. Results can vary with your photo, audio and prompt.
This example shows two people taking turns speaking while the other person reacts. It illustrates one generated result, not a guarantee that every photo, recording, or speaker handoff will work the same way.
The quality of your photo, audio, and motion instructions can affect the generated video. Use these practical tips to improve clarity and reduce unwanted movements.
A sharp, well-lit image makes it easier to see mouth movements and facial expressions. Avoid heavily obscured faces, extreme angles, and very small subjects.
Make sure your recording contains the exact words you want the person to say. Reduce background noise, check the audio duration, and listen for cut-off words before generating.
Start without a prompt to see the basic talking result. Then try one clear gesture or expression at a time. Complex sequences may not be followed completely.
Play the result with sound and watch for mismatched lip movements, unexpected facial expressions, incomplete gestures, or changes in the original character's appearance.
The presenter looks at the camera, speaks with a friendly expression, and makes a small open-hand gesture while talking.
Example instruction, not a tested result. Results may vary.For more detailed guidance, read the OmniHuman Prompt Guide.
Turn a portrait and a voice recording into a talking video for presentations, personal messages, character stories, and social content. Choose a suitable photo, then match it with the speech you want to use.
Use a clear presenter portrait and a recorded explanation to create a talking head video for a lesson, walkthrough, or product introduction. Keep the audio focused and the gestures simple.
Pair an authorized portrait with a birthday wish, thank-you message, or personal introduction. Record the exact message you want the photo to deliver.
Give a fictional character a speaking role by combining character artwork with recorded dialogue. Use optional motion prompts to guide a simple expression or gesture.
Create short speaking-portrait clips for social posts, introductions, or announcements. Use concise audio and check the result before publishing.
Find answers about photo and audio requirements, voice recordings, generation costs, video downloads, and how to get better talking photo results.
An AI talking photo generator turns a still image into a video where the subject appears to speak. This tool uses an uploaded photo and an audio recording to generate lip movements, facial expressions, and other motion.
No. This tool requires a spoken audio file. A text prompt can guide gestures or facial expressions, but it cannot turn a written script into speech.
Yes. Record your voice separately, then upload it as an MP3 or WAV file. The tool uses the recording you provide; it does not record your voice or clone it.
Supported image formats include JPG, PNG, WebP, BMP, TIFF, GIF, HEIC, and HEIF. Images must be under 30 MB, with each dimension between 300 and 6,000 pixels and an aspect ratio from 0.4 to 2.5. Audio must be MP3 or WAV, 4–300 seconds long, and under 15 MB.
Start with a clear, well-lit portrait where the face and mouth are visible. Avoid heavily blurred images, blocked faces, and extreme angles. If you want the subject to gesture, choose a photo that shows the hands.
The tool accepts audio recordings between 4 and 300 seconds. Your uploaded recording provides the speech and determines the requested performance length. Check the generated result for any timing or synchronization issues.
No. A photo and audio recording are enough to get started. Use a prompt only when you want to request a specific gesture, facial expression, or movement. Prompt instructions are not guaranteed to appear exactly as requested.
Generation costs 1 credit per second of uploaded audio, rounded up to the next whole second. The estimated cost is shown before submission. For current credit packs, see Pricing.
Yes. Sign in to submit a generation request. You can watch the public talking photo examples without signing in.
Yes. After generation is complete, preview the result and select Download to save the video.
Try a clearer portrait and a cleaner audio recording. Test the result without a motion prompt first, then add one simple action or expression. Review the mouth movement, timing, and gestures before using the video.
Only upload photos and audio you own or have permission to use. Obtain consent from the people involved before creating or sharing a video that depicts them speaking, and avoid misleading impersonation.