FAQ
Common questions about privacy, diagram types, and how captures are processed.
Does my voice audio leave the device?
When transcription runs on-device (WhisperKit or Apple Speech Analyzer), your voice audio stays on the device. A cloud transcription option is also available if you choose it.
What kinds of diagrams can it produce?
Diagrams render as one of three layouts — flow, map, or summary — plus a Mermaid type for flowcharts and sequence diagrams.
How are my thoughts classified?
Content intelligence classifies the transcript by thought type and recommends a diagram style. On-device classification covers a set of thought types, with a confidence score shown; one-tap correction trains the model.
What’s the difference between the tiers?
The free tier is genuinely useful for daily capture, with a per-day capture limit and clipboard export. Pro unlocks unlimited captures, all thought types, screen-off walking mode, on-device classification, and direct Notion/Obsidian export. See Pricing for the full breakdown.
Do Mermaid diagrams need a network connection?
Mermaid rendering in the web viewer relies on a pinned Mermaid library loaded at view time, so the viewer expects network access when displaying a Mermaid diagram.
Can I edit a transcript before the diagram is generated?
Yes — enable manual review to edit the transcript first, so the diagram reflects exactly what you meant. See Capture.