How Lumer uses AI

To summarize: Lumer shows up only when you’ve agreed, uses only what it needs, and you can switch things off at any time.

Why Lumer shows up

Lumer calls or notifies you only at times you set up together, for tasks you create or confirm. You can decline any call with one tap. Declining has no consequence and Lumer won’t follow up about it. Why a call and not a notification? Here’s our thinking.

What information the AI uses

To respond to you, the AI uses your nickname and pronouns, your onboarding answers, today’s tasks, the current conversation and, if Memory is on, your memory notes. From your conversations, Lumer also generates call summaries (including emotional tone and friction points) and, if Memory is on, inferred memory notes such as preferences and triggers. These may reveal information about ADHD, mental health or wellbeing, including things you did not expressly say, and they may be wrong.

What happens to your voice

During a call, your voice is processed in real time and converted to text so Lumer can respond. The audio is not recorded and is discarded when the call ends. We keep the text transcript and summary of the call and, if Memory is on, any memory notes.

Who processes it

We use third-party AI and speech providers to generate responses and convert speech to and from text: OpenAI (conversation), Deepgram (speech-to-text), Cartesia (text-to-speech) and LiveKit (real-time voice calls). We may also use ElevenLabs or Google for text-to-speech.

How to turn it off

• Turn off or delete memory in Settings › Personal Intelligence.
• Change or turn off calls and reminders for any task, or in Settings › Notifications.
• Withdraw microphone or notification permissions in your device settings.

AI responses can be wrong. For the full details, see our Privacy Policy and Terms of Service. For more on why Lumer works the way it does, read the Lumer Journal.