Deep-Live-Cam 2.7 Ultimate Lands With Its Biggest Upgrade Yet
Ultimate 2.7 introduces NVIDIA TensorRT acceleration, text-guided face editing through Flux Live, RTX upscaling, additional swap models, and improvements to blending and alignment. The announcement reports substantial changes to the processing engine, reduced dependencies, batch inference, asynchronous detection, and improved caching.
Summary of a DeepLiveCam news post.
Ultimate 2.7 introduces NVIDIA TensorRT acceleration, text-guided face editing through Flux Live, RTX upscaling, additional swap models, and improvements to blending and alignment. The announcement reports substantial changes to the processing engine, reduced dependencies, batch inference, asynchronous detection, and improved caching.
The release also includes a redesigned PyQt6 interface, customizable themes, a simpler Classic View, and five interface languages. Hardware support covers NVIDIA, AMD, Apple Silicon, and enterprise GPUs. Video exports run in a separate process to improve stability. Performance figures in the source are the developer's claims and depend on hardware.
Original publication: 2026-07-17. Source author: hacksider.
More from the studio
DeepLiveCam 2.7.5 Ultimate is now released!
Version 2.7.5 adds gradual face transitions for livestreams, saved image favorites, an in-app image library, and selection of individual faces from a supplied picture. The update focuses on easier access to creative tools and smoother streaming workflows.
AI Video Editor 2.0: prompt-to-cut goes local
The full editing pipeline now runs on your machine — faster cuts, zero uploads, total privacy.
DeepLiveCam 2.7.5 Preview!
This preview outlines the planned 2.7.5 release: gradual face transitions, starred images, an integrated library, and selecting faces without manually cropping pictures. The developer describes the release as preparation for version 2.8 and its planned VTuber tools.