The Voice-First
Patient Experience
CueKeep is an ambient companion, always ready on a nearby tablet or phone. There are no menus to navigate, no passwords to remember, and no keyboards to type on. Patients simply speak.
Zero Navigation
The interface is entirely conversational. Patients do not need to learn a new app, navigate confusing hierarchies, or tap through nested screens to find memories.
Always Ready
Designed as an ambient presence, CueKeep is available whenever the patient feels disoriented, needs reassurance, or simply wants to hear a familiar personal story.
Natural Pacing
Responses are slow, clear, and deliberately warm. The companion never rushes, never corrects firmly, and never asks overwhelming multi-part questions.
"When is my husband picking me up from work?"
Patient Query (Evening)
"You've worked so hard for so many years. Let's relax here for the evening. Didn't you say you loved gardening after a long day?"
CueKeep (Non-Corrective Redirection)
Therapeutic Interaction
Non-Corrective AI Behavior
Standard AI models are trained to be factual, which means they argue. When a dementia patient asks a question rooted in past confusion, correcting them often induces severe anxiety.
CueKeep is strictly programmed to lean into the emotional reality of the patient. Instead of correcting a timeline, it validates the emotion and gently redirects the conversation to a comforting, past-oriented strength documented by the caregiver.
- Prioritizes emotional reassurance over factual accuracy.
- Prevents "argument loops" common in standard voice assistants.
The Audio Engine
Ultra-Low Latency & Emotional Prosody
The success of a conversational interface relies entirely on its responsiveness. If a system pauses for 4 seconds before answering, the patient loses the thread of the conversation or believes the system is broken.
CueKeep leverages a high-speed inference pipeline to ensure sub-second response times. Furthermore, the synthetic voice models are selected specifically for warmth, patience, and stable emotional prosody—avoiding the robotic or overly energetic tones of commercial assistants.
- Sub-second audio streaming for natural conversation flow.
- Maintains patient engagement through realistic listening cues.