Bring notes into Beatrice from other sources, or export your entire graph for safekeeping.
Export all nodes and their relationships in a portable format.
On iOS, swipe from the left edge or tap the menu icon. On macOS, the sidebar is always visible.
Found in the SETTINGS section. A count badge shows how many notes will be exported.
The system share sheet opens. Save to Files, send via AirDrop, email, or any other share target. The file is named beatrice-export-YYYY-MM-DD.txt.
Beatrice exports in Markdown format, which is both human-readable and re-importable. Each node includes:
## Title).Bring notes into Beatrice from JSON, Markdown, or plain text files.
On iOS, tap the square.and.arrow.down icon in the header toolbar. On macOS, click Import in the toolbar.
The system file picker opens. Choose a .json, .md, .txt, or any text-based file.
Beatrice parses the file and shows a preview screen with: the detected format, node count, link count, and any duplicates. Each node card shows its title, content preview, tags, and links.
Nodes are committed to your graph. Duplicate nodes (matching by ID or title) are skipped automatically. After import, Beatrice auto-connects new notes to existing ones based on shared tags and text similarity.
If you previously exported from Beatrice (or created a JSON file matching the schema), Beatrice parses the full structure including IDs, types, tags, relationships (resolved by title lookup), and timestamps.
Beatrice parses ## Heading lines as node titles. Metadata lines (**Type:**, **Tags:**, **Links:**, **Created:**) are extracted. Everything between headings becomes node content.
Any text file becomes a single node. The filename (without extension) becomes the title, and the entire file content becomes the node body.
After importing, Beatrice automatically connects new notes to existing ones.
If an imported note shares tags with existing notes, a Shared Tags relationship is created. The connection weight reflects how many tags are shared.
Beatrice compares the words in each imported note against all existing notes using Jaccard similarity. Notes with 15% or more word overlap get a Semantic Similarity connection. This lower threshold (compared to the Synapse Engine) ensures imported notes integrate into your graph immediately.
If the imported file contains relationship links (by title), Beatrice resolves them against your existing graph. Links to notes that already exist are created. Links to notes that don’t exist in your graph are silently skipped.
Beatrice protects against accidentally importing the same note twice.
A node is considered a duplicate if:
Duplicate nodes are shown in the preview with a “Duplicate” badge and are dimmed. They are skipped during import — your existing data is never overwritten.