Skip to content
TypeWhisper
DE
All platform editions

Platform edition

File Memory for macOS

Bundled

Local JSON storage for workflow memories with search and built-in management.

Version
1.0.3
Plugin ID
com.typewhisper.memory.file
Requirements
TypeWhisper 0.14.0 or newer · macOS 14.0 or newer
macOS Source Code Release 1.0.3

Inside this edition

File Memory macOS settings

Features in this edition

  • Memory storage for workflows
  • Processing without uploading audio or text
  • JSON storage inside the plugin data directory
  • View, edit, and delete stored memories

Settings in this edition

  • View, edit, or delete stored memories in the add-on view

About

File Memory stores extracted memory entries in memories.json inside the plugin data directory. Storage and text search run locally and need no API key. It is a storage provider, not a complete offline transcription or memory-extraction pipeline.

Setup

  1. Enable File Memory in Settings > Integrations.
  2. Enable TypeWhisper’s memory feature and select the LLM provider and model that extract facts from transcriptions.
  3. In the File Memory settings, search, edit or delete stored entries. The storage directory is managed by TypeWhisper; this view has no folder picker.
  4. Test a transcription containing a lasting fact and check the stored entry after extraction completes.

Data processing

The extraction LLM receives the transcription text. A cloud LLM therefore sends text to its provider even though File Memory stores the resulting facts locally. Retrieved memories can also become part of later workflow prompts. Use local providers for those steps when the full workflow must remain on the device.

On macOS, extracted entries are sent to every enabled, ready memory storage provider. Review other enabled memory add-ons as well.