Blog
Building Speech To Text Tools in public
Product updates, experiments, and lessons learned while building an STT benchmarking platform.
February 2026
CER Metrics and RTFx Performance Tracking Added
Character Error Rate and Real-Time Factor metrics now available in the Results Summary table for comprehensive STT quality and performance analysis.
WER Normalization with Nvidia NeMo is Live
High-quality multilingual WER testing and normalization powered by Nvidia NeMo. Test STT/ASR models with professional-grade accuracy metrics.
Azure Fast and Short Audio models are live
Added Azure Fast Transcription and Short Audio adapters; try them and tweak native API parameters directly.
January 2026
Announcing WER Testing and AWS Transcribe Support
New WER (Word Error Rate) testing tool and AWS Transcribe provider now available. Help us improve these features with your feedback.
Why I’m building Speech To Text Tools
The story behind the product, my background in speech work, and why unbiased STT benchmarking matters.