Speech-to-Text Battle Arena
Compare transcription providers head-to-head with live streaming or batch processing
Choose Your Fighters
Select streaming providers to battle (2 per row)
Audio Settings
Provider-agnostic audio configuration
Higher = better quality, more bandwidth
PCM16 is most compatible
Mono recommended for speech
Primary language in audio
gladia
deepgram
Provider Options
gladia
Configure options
Audio encoding format
Receive partial transcripts
Audio channels (1-8)
Silence before ending segment
deepgram
Configure options
Audio encoding format
Format dates, numbers, etc.
Speaker identification
Detect "uh", "um"
"twenty" → "20"
"five meters" → "5m"
Silence threshold for splitting
Compare Accuracy
Calculate WER and CER metrics against ground truth transcripts
Real-Time Battle
Watch providers compete head-to-head with live transcription
Batch Processing
Upload hundreds of files with annotations for bulk comparison