Skip to content

Releases: gkerherve/KherveNote

KherveNote v0.28.31

Choose a tag to compare

@gkerherve gkerherve released this 06 Oct 14:39

First public release of KherveNote — live notes for lectures, trainings and talks.

  • Write on one endless page while the talk goes on: sections, key points, questions, lists, equations ($…$, $$…$$), pictures from the camera, and PDF / Word / PowerPoint documents dropped onto the page.
  • Listen (Ctrl+L): Whisper runs offline on this computer and writes what is said in the Speech panel, line by line with the time of day. The audio never leaves the machine. Notes and speech are linked by time.
  • Claude over MCP (AI ▸ Connect to Claude): Claude Desktop, Claude Code or Cursor read the note and the speech, fill in each section with what was said and your notes missed, add sections, rewrite paragraphs and export — local-only, off until you enable it, every change one Ctrl+Z.
  • Local AI (Ollama): the same filling-in, notes from the speech, rephrasing and summaries on your own computer, no cloud key.
  • Export LaTeX, a continuous PDF (one page as long as the note) or A4 pages. tectonic and the TeX files are built in: PDF export works offline.
  • KhervePDF is included: click a PDF in a note to read and annotate it there; saved annotations stay in the note.
  • Notes save themselves, keep earlier versions, and a recording survives a crash mid-talk.

Download

  • KherveNote-Setup-0.28.31.exe — installer (per user, no admin rights; opens .knote notes; includes KhervePDF)
  • KherveNote-0.28.31-portable.zip — extract and run KherveNote.exe

The first time you press Listen, the speech model downloads once (Small, about 480 MB); after that it works offline. The local AI needs Ollama — AI ▸ Set up the local AI… guides you.

The installer is not code-signed: if Windows SmartScreen warns, choose More info → Run anyway.

macOS (Apple Silicon and Intel): see the macos-v0.28.31 release.

KherveNote v0.28.31 — macOS (Apple Silicon and Intel)

Choose a tag to compare

@gkerherve gkerherve released this 06 Oct 14:47

KherveNote 0.28.31 for macOS — the same release as Windows v0.28.31, built natively for each Mac.

  • KherveNote-0.28.31-macOS-arm64.dmg — Apple Silicon (M1 and later)
  • KherveNote-0.28.31-macOS-x86_64.dmg — Intel Macs
  • KherveNote-macOS-arm64.dmg / KherveNote-macOS-x86_64.dmg — the same images under a stable name; .sha256 checksums beside them

Open the disk image and drag KherveNote to Applications. macOS 11 or later. KhervePDF is included inside the app, for the PDFs you attach to notes.

First launch: the app is ad-hoc signed, not notarised — right-click KherveNote in Applications and choose Open, then Open again. macOS then asks once for the microphone (to listen) and the camera (to take a picture of a whiteboard or slide).

What it does: write notes on one endless page while Whisper transcribes the talk offline beside them, linked by time; Claude over MCP (Claude Desktop, Claude Code, Cursor — AI ▸ Connect to Claude) or a local AI (Ollama) fills in sections from the speech and makes notes from it; export LaTeX, a continuous PDF or A4 pages — tectonic and the TeX files are built in. The speech model downloads once the first time you press Listen (Small, about 480 MB).