Clinical accuracy across devices. A diagnosis changes from Type 2 to LADA. The system tracks the correction and never reverts, even when switching from desktop to mobile.
Doctor, 42 · 8 sessions across claude, chatgpt, gemini
Scroll through to see how facts, corrections, and context build over time.
Ask a question that requires connecting facts across sessions, hosts, and time. See where each part of the answer comes from.
What happened with Sarah Lin's diagnosis?
Sarah Lin was initially diagnosed with Type 2 diabetesC in January, but follow-up labs in February revealed GAD antibodies, leading to reclassification as LADAG. Her treatment changed from metformin to basal-bolus insulinG, and she is now in stable glycemic controlC.
Sources
Four steps. No cloud. No LLM at read time. Retrieval is deterministic, so answers don't change when you switch models or when the provider ships an update.
You talk to any AI. The bridge captures structured understanding: facts, corrections, temporal order, emotional context.
Facts, corrections, temporal order written to a local SQLite file. On your device. No cloud.
You switch platforms. The bridge traverses the graph across all sessions, hosts, and time periods.
Grounded answer from specific moments. Every claim traced to a source session. No hallucination.
Each point represents a capability that requires genuine continuity, not retrieval alone.
The LADA reclassification in February is tracked as occurring AFTER the initial Type 2 diagnosis in January. The system preserves the diagnostic timeline.
Sarah Lin (patient, LADA) and Sarah Mitchell (charge nurse, night shifts) are tracked as separate people. Critical in a clinical context where confusion could be dangerous.
The Type 2 to LADA reclassification is preserved as a correction, not a replacement. Both the original diagnosis and the correction are visible with temporal context.
Sarah Lin's initial labs from Session 1 remain available through Session 8 without re-entry. The clinical history accumulates.
Answering 'What happened with Sarah Lin?' requires connecting the initial diagnosis, the reclassification, the treatment change, and the current status across 8 sessions.
Dr. Amara used Claude, ChatGPT, and Gemini across her sessions. The LADA reclassification happened on ChatGPT but is fully available when she returns to Claude.
The weight of reclassifying a diagnosis (Session 3) and the relief when the patient responds well (Session 6) are preserved as part of the clinical narrative.