How to Make a Photo Talk — Photo, Script, and Audio Guide
Prepare one clear face and the words you want it to say. This walkthrough shows the input choices, an actual speaking-video example, and the checks to make before spending credits.
Start creatingA real photo-to-speaking-video example
The portrait and spoken script below belong to this existing lesson-presenter example. Listen to the result and compare the mouth movement with the words. It demonstrates the workflow, not a guarantee of the same result for every photo.
Spoken script
Welcome to today's drawing lesson. Start with a light outline, then build your shadows in soft layers. Keep your reference nearby, and pause whenever you want to practise.From a still photo to a talking video
Prepare a clear photo
Choose a front-facing face in good light with the mouth visible. Avoid a face covered by hair, hands, or props. Use an image you have permission to animate.
Choose text or a recording
For typed words, select Script and a supported voice. For your own delivery, upload audio or record in the browser. Expressive and LTX accept audio rather than typed scripts.
Compare model and output settings
Check the selected model's input modes, supported duration, and resolution. A longer clip or a different output setting can change the credit cost.
Check the quote, then generate
Sign in, prepare your inputs, and review the credit amount before submitting. If the balance is too low, the existing recharge flow lets you add credits. Avoid submitting the same request again while it is processing.
Review the completed video
Listen for pronunciation and watch the mouth, eyes, and face throughout the clip. Download the MP4 after checking the result. A successful credit purchase unlocks original downloads.
Choose a model before choosing a credit pack
A credit pack is a balance, not a fixed number of videos. Check your model, resolution, and duration to estimate the next render. One-time credits and subscription credits have different expiry rules.
Talking Avatar · Fast
- Resolution
- Model-selected
- Supported audio length
- Up to 600 seconds
- 5-second example
- 10 credits
- 30-second example
- 60 credits
Talking Avatar · Pro
- Resolution
- 720p
- Supported audio length
- Up to 600 seconds
- 5-second example
- 40 credits
- 30-second example
- 240 credits
Talking Avatar · Expressive
- Resolution
- Model-selected
- Supported audio length
- 4–300 seconds
- 5-second example
- 70 credits
- 30-second example
- 420 credits
Music Video
- Resolution
- 480p
- Supported audio length
- Up to 600 seconds
- 5-second example
- 10 credits
- 30-second example
- 60 credits
Fast, Pro, and Music Video use five-second billing units. Expressive also rounds usage up to five-second units. These examples use native output settings, without an optional upscale.
Examples are generation-credit estimates, not past job receipts. They use the stated model, resolution, and duration. Text-to-speech can add a separate cost; the confirmed price is shown before generation.
Check models and pricing →Generation requires sign-in and enough credits. Completed videos can be downloaded as watermarked MP4 files; a successful credit purchase unlocks original downloads. Available subscription benefits follow your active plan.
Use photos, scripts, voices, and music you own or have permission to use. Review the output and your publishing platform's requirements before sharing it.
If the result needs another try
The mouth barely moves
Check whether the mouth is visible and the voice is clear. Try a better-lit, larger face and a short section of clean speech.
The timing feels wrong
Listen to the input and the completed preview. Check for long silences, background vocals, very rapid speech, or a poorly chosen audio segment before changing models.
The language sounds different
Check the script and selected voice language. Changing the interface language does not translate the words you entered. Upload your own recording if you need a specific delivery.
Photo-to-talking-video questions
Can I make a photo talk without recording my voice?
Use Script with a compatible model and available text-to-speech voice. Check the voice preview and the generation quote before submitting.
Does watching the example use credits?
No. Watching examples and preparing inputs do not spend generation credits. Creating a new video requires sign-in and sufficient credits.
Can I reuse the same face for another script?
Yes. Keep the same photo and change the text or audio. Each new generation uses credits, and the movement can vary between results.
Is this a live interactive avatar?
No. TalkingAvatarAI renders a video file from your inputs. You can preview and download it; the result does not hold a live conversation.