AI Talking Photo

The model animates the mouth, jaw and head from a single image, so the photo does most of the work. These choices make the biggest difference:

• Face the camera. A front-facing or slightly turned face works far better than a profile.
• Keep the mouth visible. No hands, microphones, masks or hair over the lips.
• Use a sharp, well-lit photo. Low-resolution or heavily filtered images blur when they move. The AI Image Upscaler or AI Image Unblur can help first.
• Leave some space around the head. A head-and-shoulders crop gives room for natural movement.
• Write the way people speak. Short sentences and punctuation give better pauses than one long line.

Illustrations, 3D characters and paintings can work too, as long as the face has clear eyes and a mouth. A good headshot from the AI Professional Headshot tool is an easy starting point for a presenter video.

Expressions matter too. A neutral or slightly smiling face gives the model the most freedom. A wide laugh or a closed mouth in the photo can make the first frames look stiff, because the model has to open or close the mouth from an unusual starting point.

You can choose from 12 voices, six female and six male, and 10 languages: English (US), English (UK), Spanish, French, German, Italian, Portuguese (Brazil), Japanese, Korean and Hindi. Write the script in the language you pick.

The script can be up to 400 characters, which is roughly 20 to 25 seconds of speech. For longer messages, split the script into parts and join the clips in an editor. Keep the same photo and voice for every part so the result feels like one continuous video.

Voices differ in pace and tone, so try two or three with a short test line before you make the full video. A friendly, warm voice suits greetings; a calm, steady voice suits lessons and product explainers.

Want the photo to speak with a specific recorded voice or your own voice? Make a short clip of the photo with Image to Video AI, then sync your audio to it with Lip Sync AI. You can create a voice from a sample with the AI Voice Cloner, but only for voices you have permission to use.

Tip: numbers, abbreviations and brand names can be read in unexpected ways. Write them the way they should be said, for example "twenty twenty-six" or "A-P-I", and punctuate long lists with commas so the voice pauses.

Pricing follows the length of the script: $0.40 per 100 characters, rounded up, with a minimum of $0.40. A 100 character greeting costs $0.40 and a full 400 character script costs $1.60. It is pay per use with no subscription, and a failed video is refunded to your balance automatically. There is no watermark.

Only animate faces you have the right to use: your own, people who agreed, or characters you created. Never use AI Talking Photo to impersonate a real person, put words in someone's mouth, deceive viewers, harass anyone or create sexual content.

If you share a talking photo publicly, it is good practice to say it was made with AI, especially when it shows a real person. Many platforms ask creators to label AI-generated or altered media.

To keep costs predictable, write the script before you upload and check its length. The price is shown before you generate, and it goes up in steps of 100 characters, so trimming a few words can save a step.