News & releases
Everything shipping across the suite.
2.7 RC1 coming!
The RC1 announcement previews cloud transformations through Decart Live, text-driven Flux Live on supported RTX hardware, TensorRT acceleration, and the ReSwapper model. It also describes five interface languages, Classic View, more precise face mapping, smoother transition timing, improved face-edge masks, and automatic projection-window sizing.
Maximize Your DeepLiveCam Experience (Movies starring you!)
This workflow post highlights real-time playback of local video files with face transformations, without first rendering an export. It also describes mapping different source faces to multiple characters during playback.
2.7 Beta released! (Windows Nvidia/AMD and Mac!)
The 2.7 beta announcement marks the project's reported 80,000 GitHub stars and a substantial rewrite of the core application. Features include live enhancement, GPEN-256 and GPEN-512 models, adjustable enhancement intensity, optimized swapping, and frame interpolation.
2.6d Released!
Version 2.6d repairs a registration problem affecting some installations. The developer recommends the update for users who encountered registration issues and says unaffected installations do not require it.
Deep-Live-Cam Nvidia 2.6b
The NVIDIA 2.6b Quickstart package is updated to address low frame rates reported after the preceding release. The announcement identifies packaging as the primary change.
We skipped 2.5... Here comes 2.6!
The planned 2.5 update is renamed 2.6 to reflect the scope of changes. The release adds OBS-based virtual-camera support, faster processing across live sessions and file outputs, camera refresh without restarting, and general stability fixes.
Teaser! Deep-Live-Cam 2.5 and plus more surprises for subscribers!
This teaser previews improved camera handling, simpler virtual-camera setup, and GPU optimizations for the planned 2.5 release. It also announces plans for a companion real-time voice application, with a large voice collection described as included in subscriptions. These are plans stated at the article's publication date.
DeepLiveCam 2.4 is now released!
Version 2.4 fixes dropdown behavior, favors GPU processing on laptops, improves Poisson blending and ear boundaries, and adjusts the InSwapper mouth mask. It also accelerates image and video previews, with this work described as groundwork for later real-time video playback.
Version 2.3d released!
Version 2.3d adds another fix for model and camera selection dropdowns, following the earlier 2.3c patch. The article also recaps 2.3b fixes for enhancement, mouth masking, and the frame-rate counter, plus the model selection and HyperSwap additions in 2.3. Users without the reported problems are told an update is optional.
VR Preview
The developer shares an early VR technology demonstration. The post explicitly describes the work as a demo that is still being optimized, rather than an available finished release.
IShowSpeed Hypes DeepLiveCam in Electrifying Comeback Stream
The site's article reports that IShowSpeed mentioned and used DeepLiveCam during a comeback livestream. It points readers to approximately 55:40 for the mention and 1:14:40 for use of version 2.3b. The article's interpretation of the streamer's reaction is commentary, rather than an independently verified endorsement.
DeepLiveCam 2.3 Released!
Version 2.3 adds a model-selection dropdown and the HyperSwap model, offering 256-by-256 face output. It also improves the face-enhancement engine, with speed gains described by the developer as hardware-dependent. The release is distributed through the Quickstart page.