Deepgram is an AI speech recognition API platform that provides real-time and batch transcription with industry-leading accuracy, low latency, and speaker diarization for enterprise and developer applications. It powers voice features in thousands of applications globally.
Category
Audio & Music
Subcategory
Speech-to-Text
Free Tier
$200 free credits for new users
Paid Plans
Pay-per-use, Growth from $4,000/yr
API Cost
$0.0043/min for Nova-2 model
API
Deepgram Trust & Security Report
Certifications, AI-training posture, and security controls, with sourced proof.
●SOC 2 Type I●SOC 2 Type II●HIPAA●GDPR●CCPA●PCI DSS●APP 8 (Australian Privacy Principles)
Verified against Deepgram's own trust and security pages. See the full report for sourced proof →
Region
Multiple regions available: US (default), EU (api.eu.deepgram.com), Australia (api.au.deepgram.com)
Trains on Data
No
Self-hostable
Yes
Building real-time voice-to-text applicationsTranscribing call center recordings at scaleCreating voice commands for applicationsDeveloping podcast transcription servicesPowering voice AI for customer support
// MORE IN SPEECH-TO-TEXT
WhisperAudio & MusicSpeech-to-TextVerified
#speech recognition#open source
AssemblyAIAudio & MusicSpeech-to-TextVerified
#speech-to-text api#transcription api
Rev AIAudio & MusicSpeech-to-TextVerified
#transcription api#speech-to-text