For developers building voice AI products, Deepgram is a speech-recognition platform providing fast, accurate speech-to-text and text-to-speech APIs, powering voice infrastructure for companies like Five9 and Telnyx.
Presence & Market Position
#70 of 115
Traditional SaaS Going Headless
▲ 28 60-day move
Measures earned, engagement-weighted share of voice across the GTM voices panel. Methodology →
Deepgram is positioning itself as the full-stack layer for voice AI, extending beyond its original speech-to-text and text-to-speech APIs into voice-agent infrastructure and on-device deployment alongside the cloud product. The conversation around it runs across conference and expo appearances, announcements for its newest speech and voice-agent models, and integration news as other platforms add it as a selectable voice engine. Practitioner discussion covers configuration detail, like tuning interruption handling in a live voice agent, and comparison contexts where it sits alongside other speech-model providers. Hiring activity in newer markets also surfaces in the mix, pointing to active geographic expansion.
Nylas
Hightouch
AssemblyAI
Explorium
MetronomeNASA, Five9, Wistia, Telnyx, Revenue.io, CallTrackingMetrics, Creditas, Jobcase
Work at Deepgram? Claim this profile to add customers and case studies.
Business Profile
Deepgram was founded in 2015 by Scott Stephenson and a co-founder whose academic research on waveform analysis for a dark matter detector led them toward deep learning for audio, and eventually toward closing the gap they saw in speech-to-text. The company operates remote-first, with staff spread across more than 20 U.S. states and five countries, and has grown to roughly 325 people. It remains independent and has raised roughly $229 million to date. Its product line has widened from transcription into text-to-speech, voice-agent infrastructure, and audio intelligence, and it has extended into new verticals through the acquisition of OfOne, bringing its voice technology into drive-thru restaurant operations.
Agent Readiness
Measures how easily your agents can build on it — API, MCP, CLI, SDK, docs depth. Methodology →
Deepgram has built out a full agent surface on top of its core speech APIs: an MCP server for connecting directly into agent clients like Claude Code and Cursor, a CLI, and a Voice Agent API for chaining speech-to-text, an LLM, and text-to-speech into one conversational loop. Agent Skills published on GitHub extend that further, giving builders reusable patterns for wiring voice into agentic workflows.
IS THIS YOUR BRAND?
Claim this profile
Fact-check your data, add context, and earn the verified badge — free, takes minutes, work email required.
Claim Deepgram →