How to make a photo talk with AI (free, step-by-step guide)

Magic Hour publishes this guide and includes its own product in the comparison. Treat the recommendations as editorial guidance from a vendor, and verify the linked first-party product details and your own output requirements before choosing a tool.

Comparison Table

Access details checked October 10, 2026: D-ID’s watermark guidance specifies a full-video trial watermark. MyHeritage’s Deep Nostalgia guide describes several free animations and a logo on unpaid results. Deep Nostalgia uses preset portrait movements; compare that task separately from creating a photo that speaks your supplied recording.

What People Use Talking Photos For

• Bringing an old family photo to life so a loved one appears to speak during a birthday celebration, reunion, or family gathering.
• Sending a personalized greeting where a photo delivers the message instead of a standard text or video.
• Creating meme videos where a famous image, painting, or historical figure appears to say something unexpected.
• Making social media content more engaging by turning a still image into a speaking character.
• Building presentations, lessons, or training materials where a headshot photo delivers narration without requiring a talking-head video.

What Is a Talking Photo?

A talking photo tool animates a still image so the person appears to speak or sing along with audio.

Magic Hour input workflow, captured September 25, 2026. This shows the upload or prompt controls, not a completed generation. View official source

The AI analyzes facial features and generates realistic mouth and facial movements that match the uploaded sound.

Unlike a normal video, the original image never moved. The AI creates the motion from a single photo.

How Is This Different From Lip Sync?

Talking Photo starts with a still image and creates movement that was never there before. Lip Sync starts with an existing video where the person is already moving and speaking.

Use Talking Photo when you only have a photo. Use Lip Sync when you already have a video that needs to match new audio.

If you already have a video and want to sync it to new audio, Magic Hour's Lip Sync tool is what you need instead.

If you are still choosing a portrait-animation workflow, compare the best AI talking-photo tools. For moving footage, start with the best AI lip-sync tools instead.

What You Need Before You Start

Before generating your first talking photo, make sure you have:

• A clear photo of a person facing the camera
• An audio clip of someone speaking or singing
• Chrome, Safari, Firefox, or another modern browser
• A Magic Hour account for the full editor; the product page also offers a limited guest trial
• No editing software or design experience

The clearer your image and audio, the better the final result will usually look.

Prefer to Watch Instead?

Watch this quick walkthrough before following the steps below.

How to Make a Photo Talk With AI: Step-by-Step

Step 1: Go to Magic Hour's Talking Photo Tool

Open Magic Hour's Talking Photo product page for the guest trial. For the full mode, audio and resolution controls, open the signed-in Talking Photo editor. The steps below describe that editor; guest allowances are separate.

Step 2: Upload Your Photo

Select Add Image and upload a clear portrait with a visible face. Choose the source image you are authorized to animate; leave room around the head for the final crop.

Step 3: Add Your Audio

Select Add Audio → Upload from device for an existing recording. The same menu offers Text to Speech and Clone your voice with AI. Select the audio section you actually want the portrait to speak; a motion prompt does not supply the spoken words. To make the portrait speak Spanish, French or another language, upload a recording already spoken in that language. Talking Photo animates the supplied audio; it does not translate it. If starting with text, select an available text-to-speech voice and language, then review pronunciation before generating the video. See our language support guide.

Choose the mode before trimming the recording

Realistic supports up to 300 seconds; Expressive supports up to 45 seconds. Realistic has no 1080p option. Expressive can offer 1080p when the account has access. For a 90-second narration, choose Realistic or plan shorter Expressive clips when that mode and resolution are required.

If the result is silent or cut short, check the selected audio start/end and player mute state, then inspect the downloaded file. Keep the project and report the first affected timestamp if the file is still broken; avoid repeating the same input without diagnosing it. See the current Talking Photo instructions.

Step 4: Click Generate

Choose the mode, resolution and selected audio length, review the credit estimate, then select Generate Video. The permitted clip duration is not a processing-time estimate or a guarantee of accurate lip sync.

Step 5: Review and Download

Watch the completed animation to make sure the speech and facial movements look natural. If you're happy with the result, download the video to your device.

Tips for the Best Results

• Use a photo where the person is clearly facing the camera with their full face visible. Angled faces usually produce less realistic animation.
• Choose a photo with good lighting and a clean background. Clear facial details help the AI generate better movements.
• Keep your first audio clip short and easy to understand. This makes it easier to evaluate the quality of the result.
• Photos with neutral expressions often animate more naturally than photos where the person is already making an extreme facial expression.
• If the mouth movement looks slightly off, try using a higher-resolution photo and cleaner audio, then generate again.

What to Try Next

If you want to go further with your photo, Magic Hour's AI Image Editor lets you change the background, remove objects, or improve the image before animating it. For videos where a real person is already speaking and you want to sync their mouth to new audio, Magic Hour's Lip Sync tool handles that instead. Both tools follow the same simple workflow: upload, adjust, and download.

Make a talking photo survive full review

Direct answer: Use a clear portrait, clean audio and a short script. Generate a small test first, then review lip timing, teeth, blinking, identity and head motion before committing to a longer clip.

Prepare the portrait. Choose a visible face with sharp eyes and mouth, even light and limited obstruction. Leave enough space around the head for movement and the destination crop.

Prepare audio. Remove heavy noise, clipping and long silence. Use natural pacing. Test names, numbers and difficult pronunciations early. Keep the voice and likeness consent with the project.

Review. Watch at normal speed, then inspect pauses and closed-mouth sounds. Check facial drift, unnatural teeth, frozen eyes, frame edges and audio offset. Export in the final aspect ratio before approval.

Try the workflow: Open the matching Magic Hour tool. Use the same representative input and acceptance criteria before comparing results.