Deepgram is a developer-focused voice AI platform providing APIs for:
- Speech-to-Text (STT): real-time streaming and batch transcription with metadata (timestamps, diarization, formatting) and add-ons such as redaction and keyterm prompting.
- Text-to-Speech (TTS): Aura voices designed for low-latency and perceptually natural output for agents and voice applications.
- Voice Agent API: a speech-to-speech interface that integrates STT, TTS, and session orchestration to reduce integration complexity.
- Audio intelligence: post-processing tasks (summarization, topic detection, sentiment, intent) applied to transcripts or audio.
Positioning emphasizes enterprise voice AI infrastructure for builders, with options for cloud APIs and self-hosted/on-prem deployments.
Primary evidence: vendor homepage, pricing pages, and product documentation.