Why I Built Knovum: A Local-First Infinite Canvas AI Workspace

Описание и применение

Knovum – локальное решение для управления знаниями с бесконечной визуальной канвас‑средой и встроенным ИИ, обеспечивающее приватность и быстрый поиск.

Задача

Existing productivity tools force users into rigid tree directories or lock private thoughts behind proprietary cloud subscriptions.

Результат

Knovum delivers a lightweight, secure, and privacy‑preserving knowledge workspace with infinite canvas, local AI, and semantic search, outperforming traditional tree‑based tools.

Рабочий процесс

• 1. Choose Tauri 2 as the desktop foundation for low memory footprint and security.
• 2. Build the infinite canvas with Vue 3 and Konva, enabling 60 fps and JSON Canvas compatibility.
• 3. Store data locally in SQLite and index vectors with LanceDB for instant semantic search.
• 4. Transcribe audio locally using Whisper + FFmpeg and slice into time‑coded note cards.
• 5. Integrate Model Context Protocol for bi‑directional AI agent workflows.

Доказательства

Knovum stays exceptionally light and boots in milliseconds.

Источник

Full compatibility with the open JSON Canvas specification. No vendor lock‑in.

Источник

Everything is stored in local SQLite databases on your machine.

Источник

Whisper + FFmpeg: Drop audio recordings or meeting logs directly onto the canvas; they are transcribed locally and automatically sliced into time‑coded note cards.

Источник