AI Transparency Statement

Version 2.0Effective 1 June 2026

9.1 AI Features

  • Speech-to-Text Transcription: converts voice recordings to text using on-device ML models; may contain errors you can edit; can be disabled.
  • Mood Detection: analyses memories to identify emotional patterns via sentiment analysis; can be disabled.
  • Keyword Extraction: identifies key topics and themes to help you organise and find memories; reviewable and editable.
  • Recommendations: suggests memories based on your patterns; can be disabled.

9.2 How AI Works

AI processing of your memory content happens on your device. We do not use your memory content to train external AI models or for any purpose other than improving your Keepr experience.

9.3 Limitations & Accuracy

AI features are not perfect. They may misunderstand accents or background noise, misclassify emotions, extract irrelevant keywords, or make incorrect recommendations. You can always review, edit, or override AI-generated content.

9.4 Your Control

  • Enable/Disable: turn any AI feature on or off.
  • Review: see AI-generated content before it is saved.
  • Edit: modify or correct AI outputs.
  • Delete: remove AI-generated data anytime.
  • Opt-Out: disable all AI features if you prefer.

9.5 No Profiling or Automated Decisions

Keepr does not create detailed profiles of you, make automated decisions that significantly affect you, use AI for advertising, share AI insights with third parties, or train external AI models on your data.

9.6 Transparency & Accountability

We provide clear labelling of AI-generated content, explanations of how each feature works, regular updates, feedback mechanisms, and accountability for AI decisions.

9.7 Contact