← Blog
Talking avatar·by Vinícius Luiz·6 min read·

How to Create a Talking Avatar From a Photo: Complete Step-by-Step Guide

Make a photo talk with AI: the real AvataReel walkthrough, from photo to MP4. Which photo to use, what gets rejected, how to record your voice, what it costs and a ready example.

How to Create a Talking Avatar From a Photo: Complete Step-by-Step Guide

To make a photo talk on AvataReel, you upload a photo of your face, choose a voice and write the script. Your first avatar made from your photo is free. The video comes out as a vertical 9:16 MP4, with optional captions and no watermark.

See a real example. This photo and a text script became the video next to it:

Uploaded photo
Original photo of Helena, used as the only input image
Generated video
One photo and a text script became this 13-second video. Tap the video to hear it.

What you need before you start

  • A photo of yourself, facing forward, in JPG, PNG or WebP, up to 10 MB. HEIC photos, the iPhone default, will not open: export as JPG or use Take photo now.
  • Your voice (optional): a quiet place to record at least 10 seconds. If you prefer, pick a ready-made voice.
  • A script: the speech of the video. You can write it on the spot or ask the AI for help.

Do not want to use your own face? Skip to No good photo: two alternatives.

Step 1: Create your account

Go to avatareel.com/signup and sign up with your email, no card needed.

Step 2: Upload the photo

Open Avatars and tap Create with my photo. In the Your photo section, use Choose photo, Take photo now or drag the image in.

A photo that works follows four rules:

  1. A real photo, not a drawing or cartoon.
  2. Look straight at the camera, with your face level.
  3. From the top of the head to the chest, at arm's length.
  4. Facing the light, no sunglasses and only you in the frame.

In a few seconds, AvataReel checks the photo. Here is what happens with each problem:

Problem with the photoWhat happens
No face, several people, drawing or illustrationBlocks. Upload another photo
Small photo (under 480 px) or very darkWarning. You can swap it or create anyway
Small face or too close to the cameraWarning: move closer or step back with the phone
Shaky or out-of-focus photoWarning: clean the lens and hold still
Dark face or blown-out lightWarning: face a window, with no direct sun or flash
Closed eyes or looking asideWarning: look at the lens with your eyes open
Turned or tilted faceWarning: hold the phone at eye level
Hand on the face or sunglassesWarning: keep your whole face visible

When there is a warning, you see "The photo quality is poor" and choose between Use another photo and Create anyway. We recommend swapping: the photo weighs the most on the result. No video quality fixes a shaky photo.

Step 3: Choose the voice

In the Voice section, you have two options:

  • Ready-made voice: no recording needed. Listen to the samples and pick one.
  • My voice: record on the spot or upload audio in MP3, M4A, WAV, OGG or MP4, up to 25 MB.

While recording, the bar marks the 10-second minimum, enough at 45 and stops on its own at 90. AvataReel shows a Script to read, so you do not freeze.

To make the voice sound like yours:

  • Record in a quiet place, with no TV, fan or street noise in the background.
  • Phone a hand's width from your mouth, no Bluetooth headset.
  • Speak as in your videos: natural tone, unhurried and with no long pauses.
  • Read 2 or 3 scripts in a row. We use the cleanest passages.

After you create the avatar, the recording is checked in about a minute. If there is noise, distortion, muffled sound or long silences, you see Record your voice again, with the reason. Meanwhile, the avatar already creates videos with a voice similar to yours.

Step 4: Confirm the authorization and create the avatar

In the Authorization section, check that you have the right to use the image and the voice. Tap Continue, review the photo under Happy with how it looks? and tap Create free avatar.

The first avatar is free on every account. Later ones cost 15 credits, and packs and plans include extra avatars. If preparation fails, the credits come back.

When it is ready, the avatar appears under My avatars.

Step 5: Write the script and generate the video

Open the avatar and tap Use this avatar. In the video creator:

  1. Write the speech. Count 15 characters per second: 225 characters give about 15 seconds. The limit is 90 seconds of speech.
  2. No idea? Tap Help me write and choose between giving a tip, showing a product or making an invitation.
  3. Optional adjustments: Outfit changes the clothing and framing for this video only. Scene and product places the avatar in a setting and with your product photo. Captions are already on.
  4. Choose the quality: Economy (720p) or Premium (1080p). See the difference side by side in Premium or Economy.
  5. Confirm. The Confirm your video screen shows the cost before generating. If generation fails, the credits come back.

Example of a 15-second script, for a tips profile:

"Do you record video and freeze when it is time to speak? Write the script first, read it out loud once and cut everything you would not say in a conversation. Test it on your next video and tell me how it goes."

Step 6: Review and download

When the video is ready, watch it all the way through and tap Download MP4. Videos have no watermark.

Want the video more edited? Tap AI editing. It adds a title, animated captions and music, at no extra cost.

How much it costs

ItemCost
First avatar from your photoFree
Later avatars15 credits (packs and plans include extras)
15 s video6 credits on Economy, 11 on Premium
30 s video10 credits on Economy, 18 on Premium
AI editingFree

The Starter Pack has 120 credits for $9, sold by card in USD, and purchased credits do not expire. See all the prices at /pricing.

Make your photo talk

Your first avatar made from your photo is free. Videos from 4 credits.

Create my avatar →

No good photo: two alternatives

  • Catalog avatar: under Avatars, open a ready-made face, listen to the sample and tap Use this avatar. You do not need to upload anything.
  • Fictional person: when creating the avatar, tap Describe a fictional person. Write age, hair, face and clothing, tap Generate photo and wait about 1 minute. AvataReel suggests a matching voice. Only adult, fictional people, with no mention of celebrities.

Common mistakes

  • Selfie with your arm stretched to the side. The face ends up turned. Hold the phone at eye level, facing forward.
  • Photo against a window. The face ends up dark. Turn to face the light.
  • Crop too tight on the face. The model needs to see down to the chest for natural framing.
  • Recording the voice with a Bluetooth headset. The sound arrives muffled. Use the phone's microphone.
  • A script written to be read. Long sentences sound artificial. Write the way you speak.

To understand what happens between the photo and the video, read AI lip sync: how a photo becomes a talking avatar.

Start now

Create your AvataReel account and make your photo talk. If you want to understand more about the technology first, read What is an AI talking avatar or the complete guide.

Frequently asked questions

Do I need a professional photo?

No. A sharp selfie works, as long as it is facing forward, with your face straight, looking at the lens, from the top of the head to the chest and facing the light. The shorter side of the photo needs to be at least 480 px.

Which photos are rejected?

Photos with no face, with several people, or that are drawings, cartoons or illustrations are always rejected. A dark, blurry photo, with a turned face, closed eyes, a hand on the face or sunglasses gives a warning, and you can swap the photo or create anyway.

Can I use my own voice?

Yes. Choose My voice and record or upload at least 10 seconds. Recording goes up to 90 seconds, and 45 seconds is already enough. If there is noise or distortion, AvataReel asks for a new recording and tells you why.

How much does it cost?

Your first avatar made from your photo is free. Later ones cost 15 credits, and packs and plans include extra avatars. Videos are priced per second: 15 seconds costs 6 credits on Economy and 11 on Premium.

I do not have a good photo. Can I still create one?

Yes. Use a ready-made avatar from the catalog, or tap Describe a fictional person and describe age, hair, face and clothing. AvataReel generates the photo and suggests a matching voice.

How long can the video be?

Up to 90 seconds of speech. A good estimate is 15 characters per second, so 225 characters give about 15 seconds.

Make your first video with your own face

Upload a photo, record your voice and type the script. Your first avatar is free.

Create my free avatar →
How to Create a Talking Avatar From a Photo: Complete Step-by-Step Guide · AvataReel