Edge Computing

While the dedicated hardware handles secure recording, the user's smartphone serves as an edge computing node.

  1. Data Transfer - Recorded audio blocks stream to smartphone via Bluetooth.

  2. Language Detection: Detect speaking language.

  3. Speech Recognition - Generates full-text transcription.

  4. Topic Classification - Categorizes content with hierarchical tags.

  5. Preference Scoring - Analyzes personal relevance based on usage patterns.

  6. Metadata Creation - Packages context-rich information with the recording.

Privacy-Preserving Architecture

This distributed approach maintains privacy while enabling powerful features:

  • All AI processing happens on-device.

  • Only metadata hashes stored on-chain.

  • Raw data and transcripts remain under user control.

  • Optional selective sharing with specific requesters.

Last updated