
Every feature is engineered to reduce barriers, protect privacy, and deliver clinically accurate support to young Zambians.
At the heart of Aunty Mimi is Google's most advanced AI model, custom fine-tuned for Zambian SRHR contexts. With a 1 million token context window, it maintains entire conversation histories for nuanced, compassionate care delivery.
Trained on complete WHO SRHR guidelines, Zambian MoH protocols, PEPFAR treatment guidelines, and local clinical best practices.
Remembers context across entire sessions. Asks clarifying questions like a real clinician. Never gives generic advice — always contextual.
Triple-layer content safety: AI refuses to diagnose, recommends testing, and always suggests professional follow-up for clinical decisions.
Automatically calculates hours since potential exposure, assesses PEP eligibility within the 72-hour clinical window, and generates urgent facility referrals.
Our crisis detection AI monitors every conversation for signals of immediate danger — suicidal ideation, self-harm, GBV in progress, or medical emergencies. When detected, it triggers an instant handover protocol.
Average time from crisis detection to human counsellor connection.
Counsellors receive complete conversation history, risk assessment, and recommended intervention protocols automatically.
Police (990), Ambulance (991), Child Helpline (116), and Crisis Line (5000) are always one tap away.
Machine learning identifies trends and hotspots, enabling preventive outreach before crises occur.
For youth in hostile environments where accessing health information could be dangerous, Stealth Mode disguises the entire Aunty Mimi interface as a calculator app. One secret gesture returns to the health platform.
Full working calculator interface. No hint of health content. Activated via Ctrl+Shift+S or triple-tap on mobile.
Auto-clears browser history, cache, and cookies of all Aunty Mimi activity when Stealth Mode activates.
Escape key instantly closes the session, clears all data, and redirects to a neutral page (Google search).
The first NLP pipeline in Africa that translates youth street slang into clinical medical terminology across 4 Zambian languages.
When Aunty Mimi AI refers a user to a facility, it generates a unique encrypted QR code. The user simply shows this at the clinic — no name, no explanation needed. The facility scans it and instantly knows the recommended services.
QR encodes only clinical recommendation — never personal details. Expires after single use.
287 mapped facilities can scan and decode referrals. Service codes map to standard clinical pathways.
PEP referral QRs display remaining window hours. Facilities prioritize time-critical cases automatically.
A real-time operations dashboard for counsellors and administrators to monitor conversations, intervene in crises, and analyze health trends.
See all active AI conversations in real-time. Flag, annotate, or take over any session instantly.
Send targeted health alerts, medication reminders, and campaign messages to segmented user groups.
Geographic heat maps of query types, STI trend analysis, and crisis hotspot prediction using ML.
Admins, clinical supervisors, and peer counsellors each see appropriate data levels with full audit logging.
Seamless AI-to-human handover with full context, risk score, and recommended intervention pathway.
Anonymized aggregate data exports compatible with MoH DHIS2 systems for national health reporting.
50% of Zambian youth have intermittent internet. Aunty Mimi's Progressive Web App architecture ensures critical health information is always accessible — even with zero connectivity.
Core health knowledge base, emergency contacts, and facility data cached locally for instant offline access.
Users add to home screen directly — no Google Play or App Store needed. Works on any smartphone.
Compressed payloads under 200KB. Text-first UI loads on EDGE/2G connections in under 3 seconds.
Try our AI chat, explore the facility map, or learn more about how we protect your privacy.