KlingAIResearch/LivePortrait Alternatives

Pick it when the driving signal is only an audio clip rather than a video, at the cost of slower diffusion sampling and no retargeting controls.

Hallo: Hierarchical Audio-Driven Visual Synthesis for Portrait Image Animation

Choose it when motion should be driven by a short text prompt over a clicked region instead of a full driving video.

[AAAI 2025] Follow-Your-Click: This repo is the official implementation of "Follow-Your-Click: Open-domain Regional Image Animation via Short Prompts"

Use it for full-body character video driven by pose/3D spatial decomposition rather than close-up facial motion transfer.

Official implementation of "MIMO: Controllable Character Video Synthesis with Spatial Decomposed Modeling"

Swap in for open-domain (non-face) image animation from text, accepting much heavier diffusion inference than LivePortrait's warping pipeline.

[ECCV 2024, Oral] DynamiCrafter: Animating Open-domain Images with Video Diffusion Priors

Pick it for long, pose-driven human animation from a reference image when you can afford a full video-diffusion stack.

Code for SCIS-2025 Paper "UniAnimate: Taming Unified Video Diffusion Models for Consistent Human Image Animation".

Use it when you need expressive talking-head output from speech audio and can tolerate diffusion latency.

Official implementations for paper: DreamTalk: When Expressive Talking Head Generation Meets Diffusion Probabilistic Models

Choose it for ControlNet-style motion guidance on arbitrary subjects, accepting heavier diffusion inference and no facial retargeting.

Fine-Grained Open Domain Image Animation with Motion Guidance

Prefer it for whole-body human animation with 3D parametric (SMPL) guidance rather than portrait-only expressive motion.

[ECCV 2024] Champ: Controllable and Consistent Human Image Animation with 3D Parametric Guidance

Pick it if you need real-time streaming diffusion portrait animation and can budget far more compute than LivePortrait's warping model.

[CVPR 2026] PersonaLive! : Expressive Portrait Image Animation for Live Streaming

Use it for audio-driven talking faces with a one-line CLI, but expect slower generation and less crisp eye/lip control.

[CVPR 2023] SadTalker：Learning Realistic 3D Motion Coefficients for Stylized Audio-Driven Single Image Talking Face Animation

Reference implementation for pose-driven character video; choose it for full-body consistency, though it is heavier and less real-time than LivePortrait.

Animate Anyone: Consistent and Controllable Image-to-Video Synthesis for Character Animation

Choose it when a speech audio track drives the portrait or animal, trading LivePortrait's video driving and speed for audio conditioning.

Diffusion-based Portrait and Animal Animation

Use it for high-quality pose-guided human video from a reference image, at the cost of diffusion-scale compute and no facial retargeting.

High-Quality Human Motion Video Generation with Confidence-aware Pose Guidance

Swap in for unified audio/pose-to-human animation at 1.3B params when multi-modal conditioning matters more than inference speed.

[AAAI 2026] EchoMimicV3: 1.3B Parameters are All You Need for Unified Multi-Modal and Multi-Task Human Animation

An older keypoint-warping research codebase: fine for reproducing motion-transfer baselines, but weaker identity preservation and no retargeting/stitching.

Code for Motion Representations for Articulated Animation paper

The previous-generation keypoint-warping animator: still a valid swap for quick research baselines but visibly lower fidelity and no animals/retargeting.

This repository contains the source code for the paper First Order Motion Model for Image Animation

Older spline/keypoint motion-transfer code: a valid fallback for simple driving-video animation but weaker on identity, eyes and animals.

[CVPR 2022] Thin-Plate Spline Motion Model for Image Animation.