1. Upload one portrait
Choose a JPG, PNG, or WebP file up to 10MB. The server validates the file, checks that the image is readable, and normalizes it to JPEG. Use a clearly visible face for the best chance of a good result.
Loading…
Essential storage keeps your sign-in and preferences working. With your permission, our analytics providers help us measure page and conversion performance; advertising storage stays off.
Loading…
Product workflow
TalkingPhoto takes two user-provided inputs—a portrait and speech audio—and submits them to an AI talking-head model. This page explains the actual production path and its limits.
Choose a JPG, PNG, or WebP file up to 10MB. The server validates the file, checks that the image is readable, and normalizes it to JPEG. Use a clearly visible face for the best chance of a good result.
Provide an MP3, WAV, or M4A recording up to 10MB. Use audio you recorded or otherwise have permission to use.
One credit is reserved and the job is sent to the AI provider asynchronously. The status page reports queued, processing, completed, or failed.
After the provider returns a video, TalkingPhoto stores it in object storage and makes it available to the signed-in owner.
The current upload validation checks file integrity and image readability; it does not promise automatic face detection before generation. Start with these conditions for a better result:
Do not impersonate people, create deceptive media, or upload a person’s photo or voice without the required permission.