TranscriptAI
AI/ML SaaS ApplicationA full-featured AI transcription platform built in partnership with Pima Web Design. TranscriptAI converts audio and video files into accurate, speaker-labeled transcripts.
Technologies Used
Application Screenshots
Drag and drop upload with security controls
Model selection and speaker detection settings
Audio waveform with speaker-labeled transcript
Usage stats and account management
Project Goals
Build a production-grade transcription product that supports multiple media formats, speaker identification, fast processing, and export-ready outputs.
My Solution
I engineered an audio processing pipeline using FFmpeg for format conversion, enhancement for cleaner signals, diarization for speaker mapping, and Whisper for high-accuracy transcription.
Key Features
Results & Impact
TranscriptAI achieved 98.7 percent average accuracy on clear audio, supported up to 10 speakers, and delivered usable outputs through an interface accessible to non-technical users.