Great Care Begins Where Memory Never Ends.
A persistent AI memory assistant for doctors, powered by Cognee Cloud's Knowledge Graph.
Doctors see dozens of patients every day. Important context—past medications, allergies, lab results, and diagnoses—often becomes fragmented across documents and visits.
Traditional AI retrieves matching text but doesn't truly remember relationships between facts.
MediMem AI gives every doctor a persistent AI memory for every patient using Cognee Cloud's Knowledge Graph.
flowchart LR
subgraph Frontend["Frontend (Next.js)"]
A[Doctor Login]
B[Patients]
C[Upload Documents]
D[Memory Hub]
E[AI Doctor]
F[Timeline]
G[Mind Map]
H[Pre-Visit Brief]
end
subgraph Backend["FastAPI"]
I[REST APIs]
end
subgraph Cognee["Cognee Cloud"]
J["remember()"]
K[(Knowledge Graph)]
L["recall()"]
M["improve()"]
N["forget()"]
end
subgraph LLM["Groq"]
O[Grounded Response]
end
A --> I
B --> I
C --> I
D --> I
E --> I
F --> I
G --> I
H --> I
I --> J
J --> K
E --> L
L --> K
K --> O
O --> M
M --> K
N --> K
| API | Purpose |
|---|---|
| remember() | Build patient knowledge graph |
| recall() | Retrieve connected information |
| improve() | Strengthen memory over time |
| forget() | Permanently remove patient memory |
- AI Doctor - ask clinical questions in plain language, grounded in real history
- Upload Documents - prescriptions & lab reports become connected memory
- Memory Hub - documents, nodes, relationships, and alerts in one view
- Timeline - a patient's full history in chronological order
- Mind Map - an interactive knowledge graph, verifiable on Cognee Cloud
- Pre-Visit Brief - a one-click five-point clinical summary
- Alerts - automatic drug-conflict and allergy flags
| Layer | Technology |
|---|---|
| Frontend | Next.js 14 · TypeScript · Tailwind CSS |
| Backend | FastAPI · Python |
| Memory | Cognee Cloud (V2 API) |
| LLM | Groq (also supports OpenAI, Anthropic, Mistral, Together, Custom) |
| Deployment | Vercel (frontend) · Railway (backend) |
WeMakeDevs Hackathon - Cognee Cloud Track
