Model · Talking presenters, Video

SadTalker

A free, downloadable research model that makes a single portrait photo talk in time with an audio clip. Licensed Apache 2.0, so commercial use is allowed. Needs a graphics card and some technical setup; unchanged since 2023.

Go to SadTalker’s website

At a glance

Open or closed: Fully open
Weights and code downloadable under a standard open licence
Code and weights downloadable; README says licence changed to Apache 2.0 and the non-commercial restriction was removed.
source · checked 12 Oct 2026 · checked on the maker's own page
Licence: Apache 2.0 (Permissive (e.g. Apache 2.0, MIT))
From the README; the LICENSE file itself was not read. README disclaimer bans fraud and infringing people's portrait rights.
source · checked 12 Oct 2026 · checked, but the page was unclear or partial
Can I sell what I make?: Yes
Apache 2.0 with the earlier non-commercial restriction removed, per README. Using Tencent names or logos needs permission.
source · checked 12 Oct 2026 · checked, but the page was unclear or partial
Cost: Free option: Free to download; you pay for your own hardware or a host.
Plans below.
How you use it: On your own computer, Through another service, Inside another tool
Local Gradio app, Stable Diffusion WebUI extension, Colab, Hugging Face Space and Replicate listed.
source · checked 12 Oct 2026 · checked, but the page was unclear or partial
Skill needed: Technical
source · checked 12 Oct 2026 · checked, but the page was unclear or partial
Made by: Xi'an Jiaotong University, Tencent AI Lab and Ant Group (research release)
README lists these institutions and says it is not an official Tencent product. No country stated.
source · checked 12 Oct 2026 · checked, but the page was unclear or partial
Status: Available
Still downloadable; latest dated changelog entry is 2023-06-12.
source · checked 12 Oct 2026 · checked, but the page was unclear or partial

What you can make with it

A talking presenter, a talking version of a photo (main use)
A video, lips that match new audio

audio to videoimage to video

Input: a portrait image (or video) and an audio clip; output: MP4.

source · checked 12 Oct 2026 · checked on the maker's own page

Plans and prices

No paid plans listed.

source · checked 12 Oct 2026 · checked on the maker's own page

Rights and safety

Watermark: No watermark
A logo watermark was added in v0.0.2 and later removed, per the changelog.
source · checked 12 Oct 2026 · checked, but the page was unclear or partial
Real faces and voices: Copies a real face (any portrait photo), no consent step
Animates any face photo; no consent step. The disclaimer forbids infringing portrait rights.
source · checked 12 Oct 2026 · checked, but the page was unclear or partial

Where you can run it

Hugging Face, Replicate, Stable Diffusion web UI (AUTOMATIC1111)

Other options for “A talking presenter, a talking version of a photo”

Akool, D-ID, DeeVid AI, DomoAI, Dreamina, Dzine, EchoMimic, Elai.io, fal, FantasyTalking, Hallo2, Hedra, HeyGen, HunyuanVideo-Avatar, InfiniteTalk, Jogg AI, LivePortrait, Magic Hour, MuseTalk, OmniHuman, VEED, Vidnoz AI, Vozo, WaveSpeedAI

About this entry

Facts last checked 12 Oct 2026 against the maker’s own pages (each fact links to its source). Entry status: draft. Plain-English summary written by Plainly.