A still portrait can become a short spoken clip when you already have a clear face and a script. The free AI talking photo tool on CanvaPro animates a photo you upload, then pairs the motion with either typed text or a short audio file. You do not need an account, and the usual download does not carry a watermark from the page.
The live form asks for a portrait in JPG, PNG, or WEBP, up to 30 MB, with the face clearly visible. You then choose Input text or Upload audio, complete the security check, and click Generate Talking Photo. The page does not play a preview. Jobs are queued and often take a few minutes. When the file is ready, you download it and review the clip on your device.
What works as a source photo
A front-facing portrait with even light and an unobstructed face gives the model a much better chance of natural lip motion. Sunglasses, heavy shadows, group shots, or a tiny face at the edge of the frame tend to fail. Crop so the person fills a generous portion of the image before you upload. A phone selfie can work if the head is sharp and facing the camera.
Text mode and audio mode
Input text is the default path. You type the line the portrait should say, up to 1,000 characters, then choose a language and a voice. The language list is English, Indonesian, Japanese, Spanish, Korean, Portuguese, and French. English voices include Lovely Girl, Wise Lady, Playful Girl, Trustworthy Man, Deep-Voiced Man, and Patient Man. Other languages have their own named voices on the form.
You can also listen to short samples on the page: Branded (2 seconds), Business (3 seconds), and Creative (4 seconds). Use those to hear the style before you commit to a longer line.
Upload audio is the other tab. It accepts an MP3, M4A, WAV, or OGG file that is under 5 seconds long. Keep the recording short and clear. If you already have the exact take you want the lips to follow, this is the better path. If you only have a written script, stay in text mode and pick a voice.
A scene you describe in words, with no portrait to animate, belongs on the text to video page instead of this talking-photo form.
How to make a talking photo, step by step
- Get consent first if the face belongs to someone else. Do not animate a person who has not agreed to it.
- Upload a clear portrait in JPG, PNG, or WEBP, 30 MB or smaller, with the face clearly visible.
- Choose Input text or Upload audio. In text mode, write the line, pick a language, and pick a voice. In audio mode, add a clip under 5 seconds.
- Complete the security check and click Generate Talking Photo once. Leave the tab open while the job sits in the queue.
- Download the clip when it is ready and watch the mouth, eyes, and head motion at full size. Keep the original still as your master photo.
Upload a clear face, add a script or audio you have the right to use, and preview the mouth before you download; a second try on the talking photo tool should start from the same still.
Good uses, and a clear limit
Creators use talking photos for short greetings, product explainers, and social hooks when they have a still but not a filmed take. Educators turn a portrait into a brief spoken prompt. Marketers test a spoken line before they book a shoot. In every case, treat the result as synthetic media and only animate faces you have permission to use.
This is not a substitute for a recorded interview, and it will not convincingly replace a long speech. Keep the line short. A few seconds of audio or a compact sentence usually looks more natural than a paragraph.
If the still should simply gain camera motion, not a spoken mouth, turn the photo into a short video instead.
If the clip fails or the mouth looks off
Recheck the file type and the 30 MB limit, and make sure the face is unobstructed. In audio mode, confirm the clip is under 5 seconds and in a supported format. An expired security check or a busy queue can also stop a job. Try once more after you fix the input.
If you only need a still image to move, without spoken audio, the image to video tool is a better match. That path animates a photograph with an optional motion prompt and does not try to lip-sync a voice.
Uploads travel over HTTPS. Files are processed to complete your request and are not sold. Only upload portraits you have the right to use, and only with the subject’s consent.
A scene you describe in words, with no portrait to animate, belongs in how to generate a short video from text.
A short checklist
- Start with a sharp, front-facing portrait under 30 MB, and get consent first.
- Use Input text (up to 1,000 characters, plus a language and voice) or Upload audio (under 5 seconds).
- Click Generate Talking Photo once and wait a few minutes for the download — there is no preview.
- Open the saved clip and check the lip motion before you share it.
When you are ready, pick one clear portrait, write a short line or add a brief recording, and save the clip only after you have watched it. One consented, finished file is far more useful than a stack of unauthorized experiments.