Phokus v0.1.0
Phokus is a local-first desktop media library for Windows — point it at your image and video folders and it builds a fast, searchable gallery with thumbnails, semantic search, visual discovery, AI tagging, and duplicate cleanup. Everything is processed on your machine; nothing is uploaded.
This is the first public release. Expect rough edges, and please file issues.
Install
- Download
Phokus_0.1.0_x64-setup.exebelow (standard CPU/DirectML build — recommended for everyone). - Run it. Windows SmartScreen will warn the publisher is unrecognized — this build is not code-signed (signing is on the roadmap). Click More info → Run anyway.
- Requires Windows 10/11. The installer fetches WebView2 automatically if needed.
NVIDIA GPU build (optional)
For faster embedding / semantic-search indexing on an NVIDIA GPU, download Phokus_0.1.0_x64-cuda-setup.exe instead. It bundles the CUDA runtime, so it's a much larger download (~295 MB) and needs a recent NVIDIA driver (CUDA 12.x). Everyone else should use the standard build.
On first launch, Phokus runs a short guided tour and downloads FFmpeg in the background (one-time). AI tagging (~1.3 GB) is optional; the CLIP model (~580 MB) downloads automatically the first time embeddings run. All of this stays on your machine.
Highlights
- Virtualized gallery for very large libraries; favorites, ratings, filters, sorts
- Filename, semantic (
/s), and tag (/t) search - Similar-image search, Explore clusters + tag cloud, EXIF Timeline
- Lightbox with zoom/pan and a custom video player
- Optional on-device AI tagging (WD tagger)
- Exact-duplicate finder with bulk delete
- Built-in updater
See the changelog for the full list.
Privacy
Your media and all derived data never leave your machine. No account, no telemetry, no cloud. The only network activity is the one-time tool/model downloads and update checks against this Releases page.
Known limitations
- Unsigned installer — SmartScreen warning as above.
- Windows only for now.
- Standard build uses CPU/DirectML inference; large libraries take time to embed on first index (the CUDA build above is faster on NVIDIA GPUs).
Verify your download (optional)
SHA-256 (Phokus_0.1.0_x64-setup.exe) = 5ac01b43ed857d3e39e73fcff7fe06b66b6a7472e1f1604abb082e2ebacd508d
SHA-256 (Phokus_0.1.0_x64-cuda-setup.exe) = 47fe02679fab1f1ea67811d73977754c18ba646dfd222985de06f056fef4bfcf