Talking Photo AI
Turn one photo into a talking head video: type a script or upload audio, and the face says it.
What is talking photo AI?
Talking photo AI turns a still portrait into a talking head video: you add one photo and a voice, either a script read by an AI voice or your own MP3 or WAV, and the face moves its lips, eyes and head in time with the audio. On blubiu it works on photos of people, illustrations and 3D characters, and each clip follows the length of the audio, up to 60 seconds.
Checked 2026-11-03
Who it's for
- +Presenter clips without a camera: talking photo AI turns one portrait into a talking head video for a product update, a course intro or a sales message.
- +Your IP character or mascot needs a voice: pick it from your character library and give it a line. In our test, a clay cat mascot talked with expressions and gestures.
- +You already have a voiceover, a podcast clip or a recorded message and want a face on screen: upload the MP3 or WAV and the photo lip-syncs it.
- +You looked for an app to make photos talk but would rather not install one: talking photo AI runs in the browser on a phone or a computer.
- +An AI talking head for social posts: the same host with a new script every day. Keep one portrait and reuse it, so every clip shows the same face.
When not to use it
- −You need walking, dancing or full-body acting: a talking photo moves the face and upper body, not the whole person. Copy a movement clip with motion control AI instead. AI Motion Transfer →
- −You want your character inside existing footage, such as a trend or an ad: AI character swap puts it into the original video. AI Character Swap for Video →
- −You want a whole scene with camera moves, other people and sound effects rather than one face talking to camera: generate it as a video shot instead. Create Video with AI →
- −It's a real person's face and they haven't agreed to it. Don't make anyone appear to say something they didn't say.
How to make a photo talk with talking photo AI
- 01
Add a photo
Upload a portrait or pick a character from your library. A clear face looking at the camera, framed from the chest up, gives the cleanest mouth movement.
- 02
Give it a voice
Type the script, pick one of 10 voices, click Make voice and listen to it. Or switch to Upload audio and add your own MP3 or WAV of up to 60 seconds.
- 03
Generate the talking head video
The Generate button shows the cost, worked out from the length of the audio. In our tests a 5–7 second line took 6 to 7.5 minutes; you can close the page, and the clip appears on your Creations page.
Tips for better talking photo AI results
Start from a front-facing, chest-up photo
Talking photo AI animates what it can see: the mouth, the eyes, the head and, in a wider frame, the arms. Our tests used chest-up crops of full-length photos; the faces stayed sharp and one speaker even added a wave. Avoid sunglasses, hands over the mouth and strong side profiles.
Leave room around the face
Talking head videos move more than you'd expect: heads turn and hands come up into the frame. In one of our tests a raised hand waved in from the edge. Crop with some space above the head and beside the shoulders so the movement isn't cut off.
Give characters a clear face
Illustrations and 3D characters work when the eyes and mouth are easy to find; our clay cat mascot blinked, smiled and gestured while it talked. Design the character with the character turnaround sheet generator, then crop its front view and use that as the photo.
Listen to the voice before you make the video
The voice is a separate, cheap step billed per 100 characters, while the video is billed by the second. Make the voice, play it back and redo it until the pace and pronunciation are right. Editing the script or the voice clears the old take, so a stale one never goes into the video.
Keep each clip under a minute
One talking photo takes up to 60 seconds of audio. In our tests English ran 15 to 19 characters a second, so the 800-character script box is roughly 45 to 55 seconds. For a longer talk, split the script at natural pauses, make one clip per part and join them in your editor.
Expect a short pause at the end
The model renders in steps of about 2 seconds, so the talking head video comes out a little longer than the audio: in our tests 5.1 and 6.4 seconds of speech both became 6.6-second clips, and 6.8 seconds became 8.7. The extra time is a natural pause at the end and counts toward the billed length; trim it in your editor if you need a tight cut.
Talking photo AI examples
All of these were generated on blubiu.
Talking photo AI vs. similar tools
| Tool | You give | You get | Best for |
|---|---|---|---|
| Talking photo AI (this page) | One photo + a typed script or an MP3 / WAV | A talking head video about as long as the audio (up to 60 s), lips synced to the words | Saying exact words: presenters, characters, greetings, voiceovers |
| Image to video | One image + a prompt describing the motion | A 1–15 second clip; any sound is generated by the model, not your script | Motion and atmosphere when the exact words don't matter |
| AI character swap | A video + your character image | The same video with your character in it | Putting a character into existing footage |
| AI motion transfer | A full-body photo + a movement clip | The person in the photo copies the moves | Dances and full-body motion |
Talking photo AI FAQ
What is a talking-head video?
A talking-head video is a shot of one person, or one character, speaking to the camera, usually framed from the chest up: presenters, explainers, vlogs and video messages all use it. Talking photo AI makes one from a still image, also called a talking image, so you get the shot without filming anyone.
Is talking photo AI free online?
Yes, to try. You can make talking photos online free with the Free plan's monthly credits, which cover a short clip plus its voice. The buttons show the exact cost before you confirm: the voice is billed per 100 characters and the video per second. Failed generations are refunded automatically; plans and credit packs are on the pricing page.
Is there an app to make photos talk?
You don't need to install one. Talking photo AI on blubiu runs in the browser on a phone or a computer: upload the photo, type or upload the voice, and download the MP4 when it's done.
Does it work on cartoons, anime or mascots?
Yes, as long as the face is easy to read. Our clay cat mascot test talked, blinked and gestured. A character without a visible mouth, or one seen only from the side, gives weaker results, so give it a front view.
How long can a talking photo be?
Up to 60 seconds of audio per clip. The video comes out slightly longer than the audio, rounded up in steps of about 2 seconds with a short pause at the end, and you're billed for the video's length. For longer talks, make several clips and join them.
Can I use my own voice?
Yes. Switch to Upload audio and add an MP3 or WAV recording of up to 60 seconds; the face lip-syncs whatever you recorded. Other formats such as M4A aren't accepted yet, so export to MP3 first.
How long does it take?
In our tests, 5–7 second lines took 6 to 7.5 minutes each. You can close the page: the job keeps running and the video appears on your Creations page.
Can I make a photo of a real person talk?
Only your own photo, or one of someone who has agreed to it. You're responsible for the faces and voices you upload, and a talking photo must not be used to make anyone appear to say something they didn't say.
Do failed generations cost credits?
No. Failed, rejected or cancelled generations are refunded automatically, for both the voice and the video.
Tools to pair with talking photo AI
Consistent Character Sheet Generator →
Lock in a person's face first, then crop the front view as the photo, so every talking clip shows the same face.
AI Character Turnaround Sheet Generator →
Design an IP character or mascot as a front, side and back sheet, and use its front view as the photo that talks.
AI Outfit Changer →
Dress the same character for each occasion first, then make each outfit talk.
Create Video with AI →
Need the character moving through a scene rather than talking to camera? Generate the shot as a video from a prompt and reference images.
Other creation tools
Wire script, character, shot and voice nodes into one pipeline and generate it clip by clip, with the same character in every clip.