Technical portfolio project
InscribeAI: AI‑Powered Meeting Intelligence Platform (Transcription, Summaries, Action Items)
A cloud-ready web platform that ingests meeting audio/video and produces structured, searchable outputs: transcripts with speaker labels, agenda coverage signals, decisions/action items, and polished PDF/DOCX exports.
This entry documents MuFaw research and technical work. It is not presented as a paid client engagement unless the project page explicitly states otherwise.
Project details
What we delivered
Overview
Automates post-meeting documentation: transcription, speaker labeling, agenda tracking, decision/action extraction, and report export.
Architecture
Upload/record → secure storage → backend job → Whisper transcription → diarization → agenda matching → LLM extraction (summary/decisions/actions) → store results → React UI + transcript chatbot → exports.
Security Posture (Practical)
Encrypted transport, encryption at rest support, auth + rules-based access controls, optional PII redaction, configurable retention.
Tech Stack
Speech-to-text: Whisper
Backend: Flask (Python) + orchestration
Frontend: React
Auth/DB/Storage: Firebase
Infra: Docker + EC2 (example deployment)
Explore further