CliniQ+

Clinical conversation to reviewable record

CliniQ+ connects live consultation audio, structured clinical fields and role-specific workspaces. Deepgram supplies the transcript; Groq extraction organizes vitals, symptoms, examination, diagnoses, medication and test fields. Transcript and record remain available together for clinician review.

Next.js/TypeScript interfaces and Supabase persistence carry patient and consultation context through doctor, patient and reception workflows. Team Queue<T>π received the Best Presentation Award at MVPM National Hackathon 2026, an event with 100+ teams.

Voice to structured record

1. Audio capture
2. Deepgram transcript
3. Groq extraction
4. Structured fields
5. Doctor workspace

Transcript, extracted fields and review surfaces remain connected throughout the consultation.

Connected product features

• Clinical record: transcript capture, structured information extraction, ICD-10-oriented output and missing-field prompts in the doctor’s workspace.
• Safety Guard: prescription interaction and allergy checks, severity levels and an override workflow that records the clinician’s reason.
• Medicine comparison: branded, generic and Jan Aushadhi price context organized around the prescribed molecule.
• Patient and operations views: records, prescriptions and an EMR-context chatbot for patients; registration and queue management for reception; administrative and research-facing screens.

Engineering decisions

Interim and finalized speech remain separate. A short-lived token proxy enables browser WebSocket transcription while keeping the provider credential on the server. Extraction and missing-field prompts connect the model output to reviewable application fields. Configured safety logic checks medication interaction and allergy context; override reasons remain part of the workflow.

The localStorage queue retains consultation identity, transcript entries and synchronization state. Triggered synchronization sends extraction requests sequentially and records successes/errors. This separates local capture from cloud processing. SHA-256-linked event payloads and timestamps provide a tamper-evident audit chain whose hashes and links can be recalculated.

Role-specific entry points give doctors a consultation workspace, patients records and an EMR-context chatbot and reception registration/queue tools. Shared patient context joins those views without reducing the application to one chat screen.