Freedom251 › AI Video Lip Sync Tools
Wav2Lip
What Wav2Lip is
Wav2Lip is open-source code for matching a video's lip movements to speech. Its repository includes inference and training code plus pretrained models, and describes use with different identities, voices, and languages, including CGI faces and synthetic voices. For inference, users can provide a supported audio file alongside a video, or use a video file from which the script extracts audio. The project includes training instructions for the LRS2 dataset; training without the visual quality discriminator takes under one day, while training with it takes about two days. The code is restricted to personal, research, and non-commercial use, and commercial use of models trained on LRS2 is prohibited. For commercial work, the repository points to Sync Labs, which offers a hosted lip-sync API and says its commercial model is higher quality than the older open-source model. Sync's paid plans include API access and SDKs, with different limits and features: Hobbyist is $50.05 a month, Creator $190.05 a month, Growth $490.05 a month, and Scale $2,490.04 a month. Sync shows Python and TypeScript SDKs, and its plans include additional features such as dubbing in 29 languages and Batch API.
Who it suits
The open-source code suits people doing personal, research, or non-commercial lip-sync work who can use its inference or training materials. Commercial users should consider the separate Sync Labs offering and its paid plans.
What it does well
- Training, inference, and pretrained models are available
- Accepts audio files or extracts audio from video
- Describes support for CGI faces and synthetic voices
- Sync offers hosted API access and SDKs
What to know first
- Open-source code is limited to non-commercial use
- LRS2-trained models cannot be used commercially
- Sync plans are paid
Limits
The open-source license allows personal, research, and non-commercial use only, and prohibits commercial use of models trained on LRS2. Sync's paid tiers limit video length, concurrent jobs, voice clones, seats, and other features by plan.
Plans
| Plan | Billed and limits | Price |
|---|---|---|
| Creator | /month · Videos up to 5 min · 3 jobs at a time · 5 voice clones · Own text-to-speech API key · Active Speaker Detection · No watermark | $190.05 a month |
| Scale | /month · Videos up to 30 min · 15 jobs at a time · 50 voice clones · 20% usage discount · 5 team seats · Delegated support channel · Batch API | $2,490.04 a month |
| Growth | /month · Videos up to 10 min · 6 jobs at a time · 15 voice clones · 5% usage discount · 3 team seats · Additional seats $25/mo | $490.05 a month |
| Hobbyist | /month · Videos up to 1 min · 1 job at a time · 3 voice clones · API access + SDKs · Lipsync Studio · Community Support | $50.05 a month |
| Enterprise | Talk to Sales | On request |
As github.com lists them, read 4 Oct 2026. The maker's own page has the final word.
Getting started
- Open the Wav2Lip repository website
- Choose Linux, API, or self-hosted deployment
- Install and run the inference code with a supported audio file or video
- Use the supplied training instructions and LRS2 dataset when training models
- For commercial work, review Sync Labs' hosted API and SDK options
Verdict
Wav2Lip provides code and models for non-commercial lip-sync work, with clear license restrictions. Commercial users are directed to Sync Labs, whose listed plans have different prices and feature limits.
Questions
Is Wav2Lip free?
Yes. It is listed as free and open source, with a free plan and free trial.
What platforms can run Wav2Lip?
The listed platforms are API, Linux, and self-hosted.
What inputs does the inference script accept?
It accepts supported audio files or a video file, from which it extracts audio automatically.
Can it be used commercially?
The code is restricted to personal, research, and non-commercial use, and commercial use of LRS2-trained models is prohibited. The repository directs commercial users to Sync Labs.
Facts from Wav2Lip's own pages
- Commercial offering
- The repository directs commercial users to Sync Labs, which offers a hosted lip-sync API and says its commercial model is higher quality than the old open-source model. · github.com, 4 Oct 2026
- Company
- Sync says it is a research company working on AI video solutions and lists its location as 1700 Montgomery St, San Francisco, CA 94111. · sync.so, 4 Oct 2026
- Compatibility
- The project says it works for any identity, voice, and language, including CGI faces and synthetic voices. · github.com, 4 Oct 2026
- Developer integration
- Sync’s quickstart shows API use through Python and TypeScript SDKs, with video and audio supplied by URL. · github.com, 4 Oct 2026
- Inference inputs
- The inference script can lip-sync a video to audio from supported audio files or a video file, from which it automatically extracts audio. · github.com, 4 Oct 2026
- License limits
- The open-source code is restricted to personal, research, and non-commercial use; commercial use of its LRS2-trained models is prohibited. · github.com, 4 Oct 2026
- Open-source materials
- Training code, inference code, and pretrained models are available in the repository. · github.com, 4 Oct 2026
- Paid plan features
- The Sync pricing page lists API access and SDKs, text-to-speech and custom voices, dubbing in 29 languages, Active Speaker Detection, and Batch API among its features. · sync.so, 4 Oct 2026
- Purpose
- Wav2Lip provides code for accurately lip-syncing videos to speech, as described in the paper “A Lip Sync Expert Is All You Need for Speech to Lip Generation In the Wild.” · github.com, 4 Oct 2026
- Support
- The Hobbyist plan includes Community Support, while Scale includes a delegated support channel. · sync.so, 4 Oct 2026
- Training
- The repository provides instructions for training with the LRS2 dataset and says training without the visual quality discriminator takes under one day, while using it takes about two days. · github.com, 4 Oct 2026
Where Wav2Lip stands on Freedom251
- AI Video Lip Sync Tools#20 of 28