A talking photo turns one still portrait into a video of that face speaking. Upload a photo, add a voice recording, and Solmi moves the mouth, jaw, and head in time with the audio, so the picture says the words: a talking avatar for your channel, an AI talking head for an explainer, or a spoken intro for a release. Trim the part you want, up to 60 seconds per clip, and export in 480p or 720p; the video comes out vertical or widescreen to match the photo. Only have a script? Turn it into speech with Solmi's AI Voiceover tool, download the audio, and upload it here.
A talking photo is a still picture animated so the face speaks along with a voice track. Solmi takes one portrait and an audio file and moves the mouth, jaw, and head in time with the voice, so the person in the photo appears to say the words. It is also called a talking avatar or an AI talking head.
Upload a clear, front-facing portrait or pick one of the preset characters, add the voice as an audio file or a YouTube or TikTok link, trim the part you want, choose 480p or 720p, and press Generate. The finished video appears under your videos, ready to download.
Not on this page: it lip-syncs a voice track you supply and has no text-to-speech of its own. To start from text, write the script in Solmi's AI Voiceover tool (70+ languages, a choice of voices), download the audio, and upload it here as the voice.
Both. The lip sync is driven by the audio itself, so a spoken recording works the same way a song does: the mouth follows the words. A clean recording with little background noise gives the clearest mouth movement. To make the photo sing a song instead, use the AI singing video generator.
Talking videos are billed per second of output: 4 credits per second at 480p and 8 at 720p, so a 15-second clip costs 60 credits at 480p. A clip can be up to 60 seconds, cut from a recording of up to 10 minutes. A new account on the web starts with 120 credits and no card, about 30 seconds at 480p.
A clear, front-facing portrait with one visible face, even lighting, and nothing covering the mouth. The video takes the photo's shape: a portrait photo gives a vertical 9:16 video for TikTok, Reels, and Shorts, and a landscape photo a 16:9 video for YouTube. A soft or small photo can be sharpened first with the built-in enhancement for 5 credits.
Yes. Turn on captions before you generate, and Solmi transcribes the voice and burns the words into the video in one of six styles: Subtle, Bold, Neon, Minimal, Serif, or Karaoke.