Web capture
Nothing from the web enters your memory by itself. Every capture is one explicit gesture — a paste, a drop, a click — and you always choose the destination vault. Web capture is three doors into the same, fully governed pipeline.
Screenshots and clips for this page are being recorded. Button wording may differ slightly from your build.
The capture browser
A real browser window inside Mnemosyne, made for one thing: grabbing what you are reading into memory. It keeps its own isolated sessions (your logins persist there, and nowhere else in the app).
- Open the capture browser
From the dock, open Web Capture — or click the globe button in the gallery toolbar. On first open, the browser asks which search engine you want as default (DuckDuckGo, Brave, Startpage, Google, Bing). You can change it anytime from the toolbar.
- Browse to the page
Use the address bar — bare words search on your chosen engine. The star saves the page to your favorites. Sites you log into stay logged in the next time you open the browser.
- Capture
Click Memorize this page in the toolbar — or right-click any image and choose Memorize this image to grab just that picture.
- Choose the destination vault
The vault picker opens. Pick the vault this belongs to — and tick remember this choice if captures from the browser should always go there (hold Shift on a later capture to be asked again). Need a new home for it? + New vault creates one on the spot.
- Let the engines read it
If reading the capture needs an engine you haven't installed yet — image search needs Theia, scanned PDFs need an OCR engine — a panel offers the setup right there. Later holds for the session; the capture is safe either way.
The other two doors
- Paste on the canvas — take a screenshot anywhere, paste it on the empty plane: the vault picker opens, the image becomes memory. (Pasting in the chat keeps its own meaning: an attachment for that conversation, not a memory.)
- Drag & drop from your browser — drag an image, a PDF, a Markdown or text file from any browser onto the canvas. Mnemosyne downloads it safely, checks what it really is, then opens the same picker.
What lands where
Every capture physically lands as a file in the chosen vault's watched folder — you can see it in your file explorer — and the watcher ingests it from there like any other document. Images join the gallery; pages and PDFs join your documents. A vault only accepts file types it declares: unknown types are refused, never guessed.
Standing choices are yours to revisit — every « remember this choice » is listed in Settings, per source (screenshots, web drops, browser captures), and revocable in one click.