LivePortrait

Overview

LivePortrait is an open-source tool for animating a portrait from a still image, using motion from video, audio, text, or generation. It best suits users comfortable with local AI tooling who want fine control over face animation. Its strongest case is flexible, locally runnable portrait work; setup and animal-mode hardware restrictions make it less straightforward than a hosted app.

It also supports portrait video editing by pairing a source video with a driving video. Face animation is supported, but depth animation is not, and MP4 is the stated export format. An online demo on Hugging Face Spaces offers a way to try it without setting up the repository.

Key features

• Motion-driven animation: Drive a portrait with video, audio, text, or generation. That range suits users who want more than a fixed photo effect, while the source-and-driving-video workflow also covers portrait video edits.
• Retargeting controls: Stitching and eye- and lip-retargeting modules give users more control over facial movement than a basic animation pass.
• Animal models: Fine-tuned models animate cats, dogs, and pandas from human driving motion. This is a useful creative option, but animal mode is restricted to Linux and Windows with an NVIDIA GPU and is unsupported on macOS.
• Performance: The project reports 12.8 ms generation speed on an RTX 4090 with PyTorch. Treat that figure as tied to the named GPU and framework, not a general speed guarantee for other hardware.
• Privacy and integrations: Motion templates can speed inference and protect privacy. Community resources include FaceFusion, ComfyUI nodes, a Stable Diffusion WebUI extension, Replicate, FAL, Poe, and Hugging Face Spaces.

Pricing

LivePortrait is free and open source under the MIT License, with inference code, pretrained models, and a Gradio interface for local use. There is no paid plan to compare; the trade-off is that users supply the setup and hardware rather than buying a managed desktop application.

Local installation requires git, conda, FFmpeg, Python 3.10, and downloaded pretrained weights. Commercial use needs special care: InsightFace detection models included in the project are for non-commercial research, so commercial users must remove and replace them to comply with that license.

Platforms

Human mode is supported on Linux, Windows, and Apple Silicon macOS; Intel macOS is untested. Animal mode has narrower support: Linux or Windows with an NVIDIA GPU, not macOS. The Hugging Face Spaces demo provides web access, while local inference is for users prepared to install the dependencies and models.

Who it's for

LivePortrait fits technically comfortable creators who want a free, open-source portrait animation workflow with facial retargeting controls, local inference, and several motion inputs. It is a weaker fit for users who need a turnkey installation, broad animal-mode hardware support, or commercial use without replacing the included detection models. Anyone animating a real person's likeness should take consent seriously: the maintainers warn about deepfake misuse, note that results may contain visual artifacts, and disclaim legal responsibility for generated output.

Pros and cons

• Pro: Free, MIT-licensed implementation with code and pretrained models gives users local access without a subscription.
• Pro: Video, audio, text, and generated motion inputs plus eye and lip retargeting support a more controlled workflow than a single-purpose photo effect.
• Pro: A web demo lets users explore the project without first completing the local setup.
• Con: Local use depends on several installed tools, Python 3.10, and downloaded weights, which raises the setup burden.
• Con: Animal animation excludes macOS and requires an NVIDIA GPU on Linux or Windows.
• Con: Commercial users must replace the included InsightFace detection models, so the MIT license alone does not clear every component for commercial use.

Alternatives

For a broader comparison, browse Photo Animation Software or AI Photo Animation Tools.

• Plotaverse is a freemium option with Android and iOS apps as well as macOS and web access; choose it if you want a mobile app rather than LivePortrait's local repository workflow.
• Deep Nostalgia offers several free animations and works on Android, iOS, and web; it is a simpler alternative for general photo animations.
• Hedra offers web and API access with monthly credit plans, including Basic at 20.00 USD per month and Pro at 50.00 USD per month; choose it if you prefer a hosted, credit-based service.
• Parallax Studio is free and open source for Linux, macOS, and Windows; consider it as another no-subscription desktop option.
• Zoetropic is a freemium Android and iOS app with basic motion and 3D effects in its free tier; pick it for a phone-based workflow.
• Free.ai Text to Video is an API option with daily free tokens and a cap of one video clip every two days on its free plan; it is aimed at text-to-video rather than LivePortrait's portrait-animation workflow.
• Higgsfield offers web and API access with monthly credit plans; consider it if you want a hosted service with paid credit tiers.
• Immersity AI is an API-based freemium option; consider it if you want a service rather than a local repository.

Verdict

Choose LivePortrait if you want a free, locally runnable portrait animation system with multiple motion inputs and useful facial controls, and you are willing to handle installation and licensing boundaries. Look elsewhere if you need a simpler hosted workflow, want animal animation on macOS, or need commercial rights without replacing the bundled detection models.