Free AI Just Got Faster on Your Phone and PC
You can now run powerful AI locally on almost any device — even your phone
Deep Dive
llama.cpp, a GitHub repository with 125k stars and 22k forks, has a new pre-release. The release adds WebP image support via FFmpeg. Builds are available for macOS (Apple Silicon and Intel), iOS, Linux (including CPU, Vulkan, and other backends), Android, and Windows (with CUDA, Vulkan, OpenVINO, SYCL, ROCm, and more). The release also includes UI assets and attestation links.
Key Points
- Llama.cpp, a free AI tool, now runs faster on phones, Macs, and even old computers
- You can use AI offline, which saves money and keeps your data private
- New formats like WebP are now supported — useful for images and documents
Why It Matters
AI just got more private, faster, and available on any device — even offline