VideoHighlighter 0.13.2

What changed since 0.13.1

New

• Fine-tune the image model in the action trainer (actions)
• Set up Ollama on another PC; grow the context a request needs (llm)
• A checkbox per action category; ◀ double-click steps back (timeline)
• Build a dataset from footage, step by step (experiment) (teach)
• Crops as samples, Browse and Move, rules from objects found (teach)

Faster

• Shrink frames off the UI thread; track live faces every frame (preview)

Fixed

• Action points count with no actions typed; refill an empty cache (actions)
• Name the track states one per line (tracking)
• Two planned crops are never merged into one (crop)
• On-demand runs use the detector you picked; Run Rules runs (analysis)

4 internal commits (docs, tests, CI, release) not listed.

Download

Install VideoHighlighter for Windows

One file. Run it, click through, done. It installs the app with
PyTorch for the processor, and downloads NVIDIA GPU acceleration
(ticked when an NVIDIA card is found) and the visual search model
during setup. Installs per user, no admin rights needed.

	Installer	Portable
Windows (desktop app)	Setup .exe	.7z.001
Windows (web UI)	—	.7z.001 .7z.002
macOS	—	.dmg

The installer is not signed yet, so Windows SmartScreen shows
"Windows protected your PC" on first run — More info → Run anyway.

The portable download runs on the processor out of the box; the
app offers NVIDIA acceleration and visual search as downloads the
first time you pick them. If it comes in several .7z.0NN parts,
put them all in one folder and extract the .001 with 7-Zip.

The web UI is the Tauri shell around the React frontend with the
Python engine bundled; it has no installer because no installer can
hold a 1.7 GB sidecar.

00-VideoHighlighter-Windows-Setup.zip is the older script-based
installer, kept as a fallback.