Guide

How to make a photo talkwith AI, step by step

To make a photo talk you need two things: a clear photo with one face, and something for it to say, either a typed script that an AI voice reads or your own audio. A lip sync tool then moves the mouth and head with the speech. The tools differ on your own voice, length, free use and price, so this guide shows the steps and what each one can and cannot do.

With AIOn CapCutWith your own voiceFree optionsAnd move

Updated

Quick answer

How do you make a photo talk?

Upload one clear photo with one face to a lip sync tool, give it a typed script or your own audio, and generate. In oVideo that tool is Live Avatar: a photo or one of 8 ready characters, up to 2,000 characters of script, a voice from a list of 96, and a square MP4 of up to 90 seconds with no watermark, on the Creator or Pro plan.

Two things decide the tool. If it must be your own voice, use one that accepts audio, such as CapCut, HeyGen, D-ID or Hedra, because oVideo has no audio upload or voice cloning. If the whole picture should move, not just the face, use image to video instead of a talking photo.

The short answer

Making a photo talk: what to know

What you need
One clear photo with the face toward the camera, plus a typed script or an audio clip
Typed script
oVideo, CapCut, HeyGen and D-ID read typed text with an AI voice. Hedra Character 3 needs audio
Your own voice
Upload audio in CapCut, HeyGen, D-ID or Hedra. Not possible in oVideo today
Free
HeyGen: 3 watermarked videos a month. D-ID: 14 day trial. oVideo and Hedra: paid only
Length
oVideo 90 s, Hedra Character 3 10 min, HeyGen and D-ID up to 30 min on paid plans
Movement
In oVideo the head and face move; the body and background stay still

oVideo facts checked in the app on November 6, 2026. Other tools as stated on their own sites on October 29, 2026, and in our talking photo apps roundup on November 4, 2026.

Method 1

Make a photo talk in oVideo

In the browser, from a typed script. Live Avatar is on the Creator plan ($40 a month, $30 billed yearly, 2,000 credits) and Pro, and costs about 2.2 credits per second: about 23 credits for 10 seconds, 199 for 90.

  1. 01

    Add the photo

    Open Live Avatar, upload a JPG, PNG or WebP photo with one face, or pick one of 8 ready characters.

  2. 02

    Type the script

    Write up to 2,000 characters (about 90 seconds of speech) and pick a voice. Play the previews first.

  3. 03

    Generate

    Press Generate Video. You get a 1024x1024 square MP4 with the lips in sync and no watermark.

Method 2

Make a photo talk on CapCut

CapCut’s lip sync page (checked October 2026) describes these steps in its desktop app. Its avatar page also says you can upload a photo and turn it into a talking character with an AI avatar.

  1. 01

    Import the photo

    Open CapCut on desktop, create a project and import your image.

  2. 02

    Turn on Lip sync

    In the Video tab, open Basic and enable Lip sync. Type the text and pick a voice, or click Add audio.

  3. 03

    Export

    Choose resolution and format in the export settings and save the video.

Compared

What each method can and cannot do

Read from each company’s own site on October 29, 2026. Prices are in US dollars and can change; check the tool before you buy.

ToolInputYour own voice or audioLengthFree usePaid from
oVideo Live AvatarOne photo or one of 8 ready characters, plus a typed script up to 2,000 charactersNo audio upload and no voice cloning. Voices from a list, tagged in 7 languagesUp to 90 s, square 1024x1024No, Creator or Pro planCreator $40/mo ($30 billed yearly), about 2.2 credits per second
CapCut Lip sync (desktop app)An image or a video, plus typed text or an audio clipAdd your own audio, or a custom voice from an uploaded voice sampleNot stated on its lip sync pageIts pages present the lip sync and avatar tools as freeNot stated on those pages
HeyGenPhoto avatars (up to 3 on Free, unlimited on paid) or 500+ ready presentersAudio input on every plan; 1 voice clone on Free, more on paid plansUp to 1 min on Free, 30 min on Creator and Pro3 videos a month with a watermarkCreator $29/mo ($24 billed yearly)
D-IDPhoto and video avatars, 100+ stock avatarsUpload audio on every plan; 1 voice clone from ProUp to 30 min per video14 day trial with 3 minutes and a full screen watermarkLite $5.90/mo ($4.70 billed yearly, D-ID watermark); Pro $29/mo ($16 billed yearly)
Hedra Character 3A portrait plus an audio track (audio is required)Brings your own audio by design; its API price list shows voice cloning as freeUp to 10 min, 540p to 1080p, square or verticalNo free plan listed on its pricing pageStudio from $20/mo (2,000 credits); API from 2.5¢ per second at 540p

Your own voice

Can I make a photo talk with my own voice?

There are two ways to give a photo your voice. You can record yourself and upload the audio, and the tool moves the lips to your recording. Or you can clone your voice from a sample, and the tool reads any text you type in a copy of it. Both need a tool that accepts audio.

Among the tools compared above, as their sites stated on October 29, 2026: CapCut lets you add your own audio or a custom voice from an uploaded sample, HeyGen takes audio on every plan and allows 1 voice clone on Free, D-ID takes uploaded audio on every plan and adds a voice clone from Pro, and Hedra Character 3 always works from an audio track.

oVideo cannot do either today. Live Avatar has no audio upload and no voice cloning: the photo speaks your typed script in one of 96 voices from a list. You can filter the list by language and gender, search it by name or style and play a preview of each voice, so you can pick one close to the voice you have in mind.

Talk and move

How do I make a photo talk and move?

A talking photo animates the face. In oVideo the lips follow every word and the face and head move with the speech, while the body, the hands and the background stay as they are in the photo. That is what most people mean by a photo that talks.

If you want the whole picture to move, such as a person who turns, walks or waves, or a camera that pushes in, you need image to video instead. In oVideo that is the AI Video tool: upload the photo, describe the motion and pick one of 10 video models. A clip runs 3 to 12 seconds depending on the model, for example 55 credits for 5 seconds, and it works on every paid plan from Starter.

The two tools do different jobs. Image to video follows a motion prompt and is not made to read your script word for word with lip sync; Live Avatar reads the script but keeps the scene still. For a message, use Live Avatar. For a moving shot, use image to video.

On your phone

How do I make a photo talk on iPhone or Android?

oVideo runs in the phone’s browser, so there is nothing to install. Open ovideo.io in Safari or Chrome, sign in, open Live Avatar and tap the upload box to pick a photo from your camera roll. Type or paste the script, pick a voice and press Generate Video. The finished MP4 is saved to your oVideo gallery, where you can download it.

For CapCut, the lip sync steps on its own site (checked October 2026) are written for its desktop app, so check the phone app before you plan around it.

A selfie works as the photo. Hold the phone at eye level, face a window so the light falls on your face, and keep hands and hair away from the mouth.

Price

What does it cost to make a photo talk?

In oVideo the price follows the length. Live Avatar counts your words at 2.5 a second, up to 90 seconds, and adds a small amount per character, which comes to about 2.2 credits per second: about 23 credits for 10 seconds, 66 for 30, 132 for 60 and 199 for 90. The exact price shows on the Generate Video button before you create, and if a video fails, the held credits go back automatically.

Live Avatar is on the Creator plan ($40 a month or $30 billed yearly, 2,000 credits, about 85 short talking videos) and the Pro plan ($120 a month or $90 billed yearly, 7,000 credits, about 300). Starter and free accounts cannot use it. Where the free plan is available, free accounts get other things instead: 2 AI images, 3 subtitle videos, 3 split screen videos, a Shorts preview and a Story storyboard preview.

Other tools let you try before paying (October 2026): HeyGen’s Free plan makes 3 watermarked videos a month of up to 1 minute, D-ID has a 14 day trial with 3 minutes and a full screen watermark, and CapCut presents its lip sync tool as free.

In your words

What people make photos say, and what happens

Scripts people typed into oVideo Live Avatar, shortened and anonymised. Most run 3 to 20 seconds.

“Hi, I’m a pediatrician, and I’m here to help you with your child’s growth.”
A doctor’s introduction for a clinic page or a profile. Upload a clear head and shoulders photo, paste the lines and pick a calm voice. Intros like this run about 10 to 15 seconds, so about 23 to 35 credits.
“Hello, how are you? I miss you.”
A short message for a family member. These 7 words make about 3 seconds of video for 7 credits. Use your own photo, or the photo of someone who agreed to send it.
“[Brand] is a platform that helps creators…”
A brand intro read by a presenter. If nobody wants to be on camera, pick one of the 8 ready characters instead of a photo. The video is square, so place it on a branded canvas in your editor for other formats.
“A speech for our national day of remembrance.”
A longer, formal script. Keep it under about 225 words so it fits the 90 second limit, pick a slow, steady voice and play its preview first. For a longer speech, split it into parts and join the clips in an editor.

Script starters

Six scripts to copy, with their price

Change the names and use them in any tool. The seconds and credits are what oVideo Live Avatar charges for these exact words: it counts 2.5 words a second and adds a small amount per character.

  • A birthday wish: 24 words, about 10 s, 22 credits. “Happy birthday, Emma! Another year older and still the brightest person I know. Have the best day, and save me a slice of cake.”
  • A good morning: 21 words, about 9 s, 20 credits. “Good morning, everyone. Coffee is ready, the sun is out and I am already thinking of you. Have a great day.”
  • A practice intro: 33 words, about 14 s, 30 credits. “Hi, I’m Dr. Lee. I’m a pediatrician, and I’m here to help you with your child’s growth, from the first steps to the first day of school. Book a visit and let’s talk.”
  • A note to mom: 29 words, about 12 s, 26 credits. “Hello, Mom. How are you? I miss you. I saw this photo today and wanted to tell you how much you mean to me. Call me when you can.”
  • A channel welcome: 25 words, about 10 s, 22 credits. “Welcome to our channel. Every week we post one short lesson you can use the same day. Subscribe so you don’t miss the next one.”
  • A holiday greeting: 24 words, about 10 s, 22 credits. “Merry Christmas from all of us! Thank you for being part of our year. We wish you a warm, happy and peaceful new year.”

Any tool

Tips for a photo that talks naturally

  • Use a sharp, well lit photo with one face looking toward the camera.
  • Keep hair, hands and objects away from the mouth.
  • Write the way people speak: short sentences sound more natural.
  • Only use photos of yourself, or of people who agreed to it.

Step by step

Tips for each stage

Before you start

  • Decide first whether it must be your own voice: that one choice rules tools in or out.
  • Read the script aloud once with a timer; that is roughly how long the video will be.
  • Check the length limit of the tool. In oVideo, 90 seconds is about 225 words.

Pick the right method

  • A message or a greeting: a talking photo. A moving scene: image to video.
  • A song: a tool that works from your own audio track, such as Hedra Character 3.
  • Longer than 90 seconds in oVideo: split the script and join the clips in an editor.

After you download

  • Burn in captions with a subtitle tool so the video works on mute.
  • For vertical feeds, set the square clip on a 9:16 canvas.
  • Keep the photo and the script, so you can remake a line later with the same face and voice.

How to choose

Which method fits what you need

Your own recorded voice
A tool that accepts audio: CapCut, HeyGen, D-ID or Hedra Character 3
A typed script, in the browser, with no watermark
oVideo Live Avatar, on the Creator or Pro plan
A video longer than 90 seconds
HeyGen or D-ID paid plans (up to 30 min) or Hedra Character 3 (up to 10 min); in oVideo, join several clips
A free try first
HeyGen Free (3 watermarked videos a month) or the D-ID trial
The whole photo moving, not just the face
Image to video, such as the AI Video tool in oVideo
A photo that sings a song
Hedra Character 3, which works from your own audio track
Captions on the talking photo
The oVideo Subtitle Generator after you download: 5 credits up to 1 minute

Answers

Frequently asked questions

How do I make a photo talk with AI?
Pick a clear photo with one face, give it something to say (typed text for an AI voice, or your own audio) and run it through a lip sync tool. In oVideo: open Live Avatar, upload the photo, type up to 2,000 characters, pick a voice and press Generate Video. The face and lips then move with every word.
Can I make a photo talk for free?
Some tools let you try. HeyGen’s Free plan makes 3 watermarked videos a month of up to 1 minute, D-ID has a 14 day trial with 3 minutes and a full screen watermark, and CapCut presents its lip sync tool as free (all checked October 2026). oVideo has no free talking photos: Live Avatar needs the Creator plan, and its videos have no watermark.
How do I make a photo talk on CapCut?
According to CapCut’s site (October 2026): open the desktop app, create a project and import your photo. In the Video tab, open Basic and turn on Lip sync. Type the text and pick a text to speech voice, or click Add audio to use your own clip. Then export.
Can I make a photo talk with my own voice?
Yes, with tools that take audio: CapCut, HeyGen, D-ID and Hedra all accept an audio clip, and HeyGen and D-ID also offer voice cloning. oVideo cannot do this today: it reads a typed script in a voice you pick from its list.
How do I make a photo talk and move?
A talking photo moves the face: in oVideo the head and face move with the speech while the body and background stay still. If you want the whole picture to move like a clip, with the camera or the body in motion, use an image to video tool instead.
What is the best app to make a photo talk?
It depends on what you need. For your own voice, use a tool that takes audio, such as CapCut, HeyGen, D-ID or Hedra. For long videos, D-ID and HeyGen paid plans allow up to 30 minutes and Hedra Character 3 up to 10. For a quick talking photo from a typed script in the browser, with voices in 7 languages and no watermark, oVideo Live Avatar.
How long can a talking photo be?
oVideo makes up to 90 seconds per video. HeyGen allows 1 minute on Free and 30 minutes on Creator and Pro, D-ID up to 30 minutes, and Hedra Character 3 up to 10 minutes, according to their sites in October 2026.
Is it OK to make someone else’s photo talk?
Only with their permission, and for a child only if you are the parent or guardian. Never make anyone appear to say something they would not say: a talking photo of a real person without consent can mislead people and cause harm.
How do I make a photo talk on iPhone?
Open ovideo.io in Safari, open Live Avatar, upload the photo from your camera roll, type the script, pick a voice and press Generate Video. It runs in the browser, so there is no app to install, and Live Avatar needs the Creator or Pro plan. CapCut’s published lip sync steps are for its desktop app.
Can I make a photo talk without a watermark?
Yes, on paid plans. oVideo Live Avatar videos never have a watermark, and HeyGen removes its watermark on paid plans. Free use usually adds one: HeyGen Free videos are watermarked, the D-ID trial has a full screen watermark and D-ID Lite carries a D-ID watermark (October 2026).
Can I make a photo sing?
Not in oVideo: Live Avatar speaks a typed script and has no audio upload, so it cannot follow a song. Hedra Character 3 makes talking or singing videos from a photo and your own audio track, according to its site as checked on November 4, 2026.
How many words can a talking photo say?
In oVideo, up to 2,000 characters and 90 seconds. Live Avatar times speech at 2.5 words a second, so 90 seconds is about 225 words, 30 seconds about 75 and 10 seconds about 25. A script that runs longer cannot be created until you shorten it.
Can I make an old family photo talk?
Yes, if the face is clear and looks toward the camera: scan or photograph the print sharply and upload it like any other photo. Faded or blurry faces give a weaker result. Do it only with the family’s agreement, and never make someone say things they would not have said.
Can I make a photo talk in Spanish, Arabic or another language?
oVideo voices are tagged in 7 languages: English, Turkish, Spanish, German, French, Arabic and Russian. Write the script in that language and pick a voice tagged for it. For far more languages, HeyGen lists 30+ on Free and 175+ languages and dialects on paid plans (checked November 4, 2026).
Can I add captions to a talking photo?
The oVideo download has no captions in the file, although a caption preview plays while you review the video. To burn captions in, upload the MP4 to the Subtitle Generator: 5 credits for a video up to 1 minute and 10 up to 5 minutes, with word by word caption styles.
Can I use a ready character instead of my own photo?
Yes. Live Avatar has 8 ready characters. Open one to see a larger preview with its gender and style, then type the script as usual. They suit brand intros and lessons where nobody wants to appear on camera.
What happens if my talking photo fails?
oVideo holds the credits when you press Generate Video and only keeps them when the video is delivered. If the video fails, the credits go back to your balance automatically and you can try again, for example with a sharper photo.

Related

Make a photo talk from a script

One photo, your text and a voice in one of 7 languages. A lip synced square video up to 90 seconds, no watermark.