Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.canthus.org/llms.txt

Use this file to discover all available pages before exploring further.

​
Overview

Research export mode is an optional, explicitly opt-in feature that allows users to export an anonymised dataset for personal research or to share with a clinician. It is separate from the standard data export feature. Standard export produces a full personal backup. Research export produces a stripped, aggregated dataset with no personally identifiable fields.
  • Opt-in only. The feature is off by default and has no passive data collection.
  • The user enables research export from Settings.
  • Consent is given per export, not as a persistent permission - each export requires a deliberate action.
  • Consent is reversible at any time by disabling the feature in Settings.
  • No data leaves the device without an explicit export action.
When the user enables research export in Settings:
“This lets you export a simplified, anonymised version of your Canthus data. Task names and exact times are removed. You can use this to share patterns with a clinician or to analyse your data yourself.”
Enable button: “Turn on research export” Disable button: “Turn off” The enable/disable state is stored in shared_preferences and is not part of the event log.

​
What is exported

Only fields classified as derived in the Analytics Event Schema are eligible for export. private and internal fields are never included.

​
Included in export

DataTransformation
Daily mana allocatedRounded to nearest 5
Daily mana usedRounded to nearest 5
Task mana costRounded to nearest integer
Cost calibration deltaIncluded as-is
Check-in tierIncluded as-is (full/quick/momentum/none)
Event dateRounded to week start (Monday)
Effort type split (physical/cognitive ratio)Included as-is
Pool engagement rateIncluded as-is

​
Never included in export

DataReason
Task titlesPersonal content
Body/mind/mood ratingsDirect health self-reports
Day type labelsUser-set personal context
Task identifiers (UUIDs)Could enable cross-referencing
Precise timestampsInfer daily patterns; replaced with week-level dates
PEM event detailsSensitive health signal

​
Anonymisation rules

  1. Strip identifiers - All UUIDs and entity IDs are replaced with sequential integers scoped to the export file. The mapping is not retained.
  2. Aggregate timestamps - All dates are rounded to the Monday of the week in which they occurred. Time-of-day is removed.
  3. Remove title fields - Task titles and any user-entered text are excluded.
  4. Round numeric fields - Mana values are rounded to the nearest 5 to reduce resolution.
  5. Re-identification check - If the export contains fewer than 14 days of data, warn the user that limited data may be easier to re-identify.

​
Export format

The export is a JSON file with the structure: 
{
  "exportedAt": "2025-W18",
  "schemaVersion": 1,
  "days": [
    {
      "week": "2025-W16",
      "allocatedMana": 60,
      "usedMana": 45,
      "checkInTier": "full",
      "effortSplit": { "physical": 0.6, "cognitive": 0.4 }
    }
  ],
  "taskCalibrations": [
    {
      "week": "2025-W15",
      "previousCost": 10,
      "newCost": 8,
      "lifetimeCompletions": 12
    }
  ]
}
The file is saved to the device’s documents directory and can be shared via the standard iOS/Android share sheet.

​
Not in scope

The following are out of scope for the initial research export implementation:
  • Automatic or scheduled export
  • Cloud sync of exported data
  • Export to third-party services
  • Aggregate data sharing with Canthus (no telemetry is collected server-side)