AI Talking Photo

A talking photo is a still portrait that comes alive and speaks. The face stays the one in the picture, but the mouth, eyes and head move with a voice line, so a profile picture, a painted portrait or a character you designed can deliver a message on camera.

On Mage it takes two steps in one account. First the voice: write what should be said and generate it with Seed Audio, which makes voices, dialogue, music and sound effects from a prompt, or use a recording you already have. Then the face: open Lipsync, add the photo and the voice, and generate. The result is an MP4 of the portrait speaking your line.

Talking photos work well for greetings, explainers, character introductions and social posts. Keep them honest: use your own photo, a character you created, or a picture of someone who said yes. Mage does not allow harmful deepfakes or non-consensual depictions of real people.

Seed Audio follows a prompt, so describe the speaker as well as the words: their age range, their pace, their mood and where they are speaking. "A warm, unhurried grandmother telling a bedtime story" reads very differently from "an excited sports commentator." Put the exact words in quotes and keep the sentences short; one or two sentences fit comfortably inside 15 seconds.

Seed Audio renders exactly the length you choose: 5, 10, 30, 60 or 120 seconds. For a talking photo, pick 5 or 10 seconds so the whole line fits in one Lipsync run with room to spare. Longer narration works too: generate it once, cut it into pieces of 15 seconds or less in any audio editor, and make a clip for each.

• Describe the voice, then quote the words
• Pick a 5 or 10 second length for one clip
• Match the voice to the age and mood of the face
• Regenerate the voice before changing the photo
• Keep each line to one idea

The best talking photos start from a clear, well-lit face that looks toward the camera, with the mouth visible and closed or relaxed. Head-and-shoulders framing works better than a full-body shot, because the face has more detail to animate. Glasses are fine; a hand or a microphone in front of the mouth is not.

If the photo is old or soft, sharpen it first with the AI image upscaler or repair it with AI photo restoration. If you would rather start from a new face, make one with the AI portrait generator and save it as a character, so the same face can speak again later.

Greetings: a birthday or holiday message from a family photo, recorded by the person in it. Explainers: a character who introduces a product, a lesson or a how-to in two sentences. Stories: a painted ancestor, a fantasy hero or a narrator portrait who opens each chapter of a series. Social posts: a mascot with a running joke, posted as a short clip.

When the face should do more than talk, take the next step: AI lip sync covers dubbing and longer narration, and the AI music video generator turns saved characters into singing, dancing performers.