DeepLiveCam
The Live AI DeepFake
138 followers
The Live AI DeepFake
138 followers
Go live as anyone or anything! Open-source, privacy-first AI tool for VTubers, creators & streamers. Stunning real-time face swaps & avatar creation, no coding needed.
Free
Launch Team
AWS Startup AdvisorPersonalized AWS guidance that meets you where you’re already building
Promoted
I’m super excited to share Deep-Live-Cam, an open-source AI tool I came across while exploring creative tech for VTubing and live streaming. Whether you want to create a full-blown virtual persona or swap faces for fun, this tool makes it effortless.
This is how Deep-Live-Cam stands out:
--> Real-time face & avatar swapping
--> Smooth performance even on modest hardware
--> Privacy-first: built-in safety & ethical use guidelines
--> Open-source for full transparency + community-driven innovation
It’s perfect for creators, VTubers, educators, streamers, or anyone who wants to explore digital identity in new, exciting ways.
Try it here: https://github.com/hacksider/Deep-Live-Cam
Bookmarking this one – maybe will use it for YouTube in the future.
P.S.: Bonus points for a detailed tutorial on installing! :)
Hehe! It's super interesting by also know it's a huge niche. Hope many of them can take the most of this tool. Wish you all the best!
Nice, saving it. What are the recommended requirements to run this locally?
How will you ensure people don't misuse it?
This looks incredibly fun and practical — the real-time swapping with performance optimizations + privacy safeguards makes Deep-Live-Cam a standout. It’s rare to see something open-source in this space that’s also this creator-friendly.
I’m especially curious how educators or virtual presenters might adapt this — beyond entertainment, it seems like a solid tool for accessibility or engagement in remote environments.
Great questions too — I’d also love to know if motion tracking via webcam or audio lip-sync is on the roadmap. And for contributors: is there a preferred stack or workflow for adding new avatar types?
Anyone tried this with OBS or in a Zoom setting yet?