TalkingPhoto frequently asked questions

TalkingPhoto combines one portrait photo with one speech recording and submits them to an AI talking-head model. If generation succeeds, the signed-in user can download the resulting MP4.

Not in the current production configuration. Upload your own MP3, WAV, or M4A speech recording. We do not show a text-to-speech control unless a real speech provider is configured.

The upload flow accepts JPG, PNG, and WebP images up to 10MB. The server validates the uploaded object and converts supported images to a consistent JPEG representation before generation.

Use a clear image with one front-facing or slightly angled face. Keep the eyes and mouth visible, avoid heavy shadows, and choose an image where the face is not tiny in the frame.

The production upload flow accepts MP3, WAV, and M4A audio up to 10MB. Upload only recordings you own or have permission to use.

There is no guaranteed fixed time. Completion depends on the external AI provider, model startup, input complexity, and queue load. The status page reports the job state instead of displaying a fabricated countdown.

The job is marked as failed. Reserved paid or subscription credits are released once. A free preview is a one-time provider attempt and is not restored after the provider has begun processing.

A new authenticated account can make one view-only preview from audio up to 5 seconds. It is kept for 24 hours, cannot be downloaded through the product, and does not require a credit card. A daily service-wide budget may pause new previews.

The active Creem products are $4.99 for 2 credits, $9.99 for 5 credits, $19.99 for 15 credits, and a $24.99 monthly Creator subscription with 60 credits per billing period. One credit covers each started 10 seconds. Taxes, if applicable, are shown at checkout.

Dashboard, status, preview, and download routes verify the signed-in user against the video owner. Media is delivered through short-lived signed storage URLs, and expired objects are removed by a protected cleanup task.

No. You must have the necessary rights and consent. Impersonation, deceptive deepfakes, fraud, harassment, and unauthorized use of public figures or private individuals are prohibited by the Acceptable Use Policy.

Yes. Clearly label AI-generated or AI-edited media when publishing it, especially when viewers could reasonably believe the depicted person actually recorded the message.