How to Make a Photo Talk: Turn Any Picture Into a Talking Video

Making a photo talk used to require an animation studio. In 2026 it takes one clear picture, a script, and a few minutes. AI models animate the face in your photo — lips, eyes, head movement, even hand gestures — and sync everything to a voice, so the person in the picture actually speaks your script.

This guide shows exactly how to do it, what kind of photo works best, and what to expect from the result.

A talking photo (also called a talking picture or photo avatar) is a video generated from a single still image. An AI model predicts how the face should move while speaking — lip shapes, blinks, micro-expressions — and renders a video where the person delivers your script with natural motion. Modern models like CraftStory’s Model 2.0 go beyond the face: they animate gestures and keep the person’s identity consistent for videos up to 5 minutes long.

• One photo. A clear, front-facing shot with a single visible face works best. Phone photos are fine.
• A voice. Either type a script (the AI generates speech in 30+ languages) or upload your own audio recording. On paid plans you can also clone your own voice so the photo speaks as you.
• An account. You can start free at app.craftstory.com — no studio, camera, or editing skills required.

Step 1 — Upload your photo. Sign in to CraftStory, open Model 2.0 (photo-to-video), and upload the picture. The app checks that a face is clearly visible — if the photo has no face or several people, it will ask for a different one.

Step 2 — Add the voice. Paste your script and pick a voice (30+ languages available), or upload an audio file if you already recorded the narration. On paid plans you can clone your own voice — record a short sample once, then reuse it in any video.

Step 3 — Choose the format. Portrait or landscape, 480p or 720p — pick what fits the destination (TikTok, Reels, and Shorts are portrait; YouTube and presentations are landscape). You can upscale the result to 1080p.

Step 4 — Generate and download. Hit generate. The AI animates the photo, syncs the lips to the audio, and adds natural gestures. Download the MP4 and publish it anywhere.

• One face, facing the camera. Group photos and profile shots are the most common reason a generation gets rejected.
• Good, even lighting. Shadows across the face reduce realism.
• Neutral expression works best. The AI adds the emotion from the speech; starting from a neutral face gives it the most room.
• Higher-resolution source photos produce sharper videos — but a normal phone photo is enough.

• Marketing and UGC-style ads — turn a founder’s or creator’s photo into a spokesperson video and iterate scripts without reshoots.
• Training and onboarding — narrated lessons with a consistent presenter.
• Product demos and explainers — a face makes them more watchable.
• Social content — talking-photo clips consistently outperform static images on TikTok, Reels, and Shorts.
• Personal projects — bring an old family photo to life or make a personalized greeting.

CraftStory is built for realism on long clips (up to 5 minutes with identity preserved), but it’s not the only option. HeyGen and Synthesia focus on enterprise avatar video with big template libraries; D-ID popularized short talking-head clips from photos. If you’re comparing platforms, see our honest breakdowns: CraftStory vs HeyGen and CraftStory vs Synthesia.

Make your first photo talk — free. Try CraftStory Model 2.0 — upload a photo, type a script, get a talking video in minutes. See pricing for plans.