Mindologi is an AI-powered virtual clinical assistant that runs automated psychological assessments using DSM-5 criteria. It mirrors a real mental health consultation—collecting personal, lifestyle, and symptom data, performing vector-based diagnostic reasoning, and producing clinical-style reports for users and healthcare providers.
Client
Healthcare / Product Build
Role
Full Stack & AI Developer
Year
2024
Timeline
End-to-end build
Industry
Mental Health / HealthTech
Automating the path from first symptoms to structured, clinician-ready insight.
Mental health consultations depend on lengthy manual assessments and DSM-based symptom screening before clinicians can reason about a diagnosis. The challenge was to build an AI system that simulates a real psychological consultation: ask dynamically from user answers, compare symptoms against DSM-5 criteria, generate structured diagnostic reports, and give providers a workspace to review cases and connect with patients.
Mindologi automates a psychiatrist-like workflow with intelligent prompting, structured capture, and medical vector search. Intake collects background (demographics, trauma, lifestyle). Symptom exploration deepens duration, triggers, and impact. DSM-5 diseases stored as vectors enable similarity search; top matches drive criteria verification with dynamic follow-ups. At 70–80% criteria confirmation, the system produces a report with disorder, match percentage, and alternatives—or advances to the next candidate until a valid fit emerges.
From AI consultation to provider-ready reports in one platform.
Multi-phase flow: Intake → Symptom exploration → Screening → Assessment → Report, with DSM-5 vector identification.
LangChain-driven, criteria-based questions that adapt to answers and screening results.
Verified symptoms, matched disorders, confidence %, next steps—shared with providers; PDF export.
Client: chat, progress, reports. Provider: verify cases, manage reports, schedule appointments.
Diseases and assessments indexed with MongoDB Vector Search; interactions logged for analytics.
Text chat plus optional Twilio voice so the AI doctor can speak with patients.
Seven layers from conversation to clinician review and notifications.
Manages user input and conversational context across phases.
Matches symptoms to DSM-5 disease embeddings in MongoDB.
Adapts follow-ups from criteria verification and screening.
Calculates symptom match percentages per disorder.
Compiles data, scores, and diagnosis into client/provider reports.
Clinicians validate reports and manage appointments.
SMS/calls via Twilio connecting clients and providers.
Initial assessment steps automated
Less clinician intake time
Dashboards (client + provider)
Accessible mental health triage
MERN, LangChain, MongoDB Vector Search, Twilio, and tools built for clinical workflows.
Dual dashboards for clients and providers with real-time chat and report views.
REST APIs, WebSockets, role-based access, and orchestration of AI flows.
REST APIs, WebSockets, role-based access, and orchestration of AI flows.
Stores diseases, DSM criteria, assessments, and vector search for symptom matching.
Conversational intake, dynamic questioning, and report generation.
Conversational intake, dynamic questioning, and report generation.
Optional voice consultations and appointment notifications.
Secure separation of client and provider experiences.
Fast frontend delivery and reliable backend hosting.
Fast frontend delivery and reliable backend hosting.
