Model · Talking presenters

Hallo2

A free, downloadable research model that animates a front-facing portrait photo to speak along with an English audio track, including long clips and an optional sharper high-resolution mode. Needs a powerful graphics card and Linux.

Go to Hallo2’s website

At a glance

Open or closed
Downloadable, with conditions
You can download and run it, but the licence limits use
Weights downloadable; code is MIT but no licence is stated for the weights, and parts build on other models with their own terms.
source · checked 12 Oct 2026 · checked, but the page was unclear or partial
Licence
MIT (code) (Permissive (e.g. Apache 2.0, MIT))
MIT shown for the code. The high-resolution feature uses modified CodeFormer under the S-Lab License 1.0. Weight terms not stated; builds on Stable Diffusion 1.5, wav2vec and AnimateDiff.
source · checked 12 Oct 2026 · inferred or from a secondary source; needs review
Can I sell what I make?
The terms don't say clearly
Code is MIT, but the weights have no stated licence and the high-resolution step must follow the S-Lab License 1.0 (its text was not read).
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.
Plans below.
How you use it
On your own computer
source · checked 12 Oct 2026 · checked on the maker's own page
Skill needed
Technical
source · checked 12 Oct 2026 · checked on the maker's own page
Made by
Fudan University Generative Vision Lab (with Baidu and Nanjing University)
Affiliations from the README; no country stated.
source · checked 12 Oct 2026 · checked, but the page was unclear or partial
Status
Available
Latest dated news 2025-01-23 (ICLR 2025 acceptance).
source · checked 12 Oct 2026 · checked, but the page was unclear or partial

What you can make with it

audio to videoimage to video

Square portrait, face 50-70% of frame, under 30 degrees rotation; English WAV audio. Showcases run up to about an hour.
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

Real faces and voices
Copies a real face (any portrait photo), no consent step
No consent step; the README's social-risk section only recommends informed consent.
source · checked 12 Oct 2026 · checked, but the page was unclear or partial

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, Hedra, HeyGen, HunyuanVideo-Avatar, InfiniteTalk, Jogg AI, LivePortrait, Magic Hour, MuseTalk, OmniHuman, SadTalker, 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.