Our Architecture
The File Processing Layer is responsible for extracting structured content from various document formats like PDFs, PPTs, and DOCs. Utilizing specialized libraries such as PDF.js, python-pptx, and docx, it parses and prepares the data for further processing.
The Speech-to-Text Layer captures real-time speech input or processes uploaded audio files. It converts spoken words into text using advanced speech recognition APIs like Google Speech-to-Text, forming the basis for note generation.
The Context-Aware Correction Layer refines transcription accuracy by incorporating domain-specific terminology extracted from uploaded files. This ensures that technical terms and jargon are correctly recognized and transcribed, improving overall note quality.
The Summarization Layer operates within a secure, sandboxed environment to generate cohesive summaries. It processes transcripts iteratively, respecting token limits, and combines them into comprehensive notes without compromising user privacy.
The Output Generation Layer provides the final structured notes, which can be exported as PDFs or accessed through an interactive chatbot. It ensures that the output is user-friendly and accessible, supporting various formats for convenience.
