Voicebuds use cloud-based AI to process your voice. This article explains how that works, who handles your data, and what you're responsible for when recording in shared spaces.
How voice activation works
When Voicebuds are powered on, the device may monitor for an activation command. When detected:
- Audio is captured and sent to Subtle's cloud for transcription
- The transcript may be processed by an AI language model
- The AI output is returned to your device
Third-party AI processing
Subtle uses enterprise AI providers — which may include companies such as OpenAI or Anthropic — to process Dictation Data. These providers:
- Are contractually prohibited from training their models on your data
- Delete your Dictation Data within 30 days
- Operate under enterprise agreements with stronger data protections than consumer AI products
Important: Subtle cannot provide absolute guarantees about third-party operations. For complete details, review Subtle's Privacy Policy.
Context Awareness (optional)
If you enable Context Awareness, Subtle may process limited surrounding context — such as the active app name or visible text — to improve formatting and accuracy of dictation. You can disable this in Settings.
Your consent responsibilities
Voicebuds can capture audio in shared environments. You are responsible for:
- Informing others in your environment that recording or transcription may be occurring
- Obtaining any necessary consents required by law in your jurisdiction before recording conversations that include other people
- Not using Subtle's services to capture sensitive personal data about others without appropriate authority
Subtle's Terms of Use place these responsibilities on the user. This is especially important in professional, legal, or healthcare settings.
Reducing data exposure
Enable Privacy Mode in Settings → Data and Privacy to minimise how long your Dictation Data is retained. See Privacy Mode: what it controls for details.