Graded C+ on resilience. In Vannus's assessment it builds its own model-layer IP on top of a third-party core — acceptable, with limits worth knowing.
Speech-to-text, text-to-speech and voice agent APIs (Nova, Flux, Aura model families) with a self-hosted deployment option on the Enterprise plan. The privacy policy names Deepgram, Inc., 548 Market St Suite 25104, San Francisco, CA and states 'We store data on servers in the U.S.' The vendor publishes a SOC 2 Type II audit. Neither the privacy policy nor the data security page addresses whether c
Deepgram builds its own model-layer IP on top of a third-party core. This is Vannus's assessment from its curated catalog intel.
This is a separate question from the grade above. The grade measures resilience — whether the tool endures and whether you could leave it. This measures who can compel your data. A tool can score modestly on one and strongly on the other, and many do.
The vendor's published or catalog-recorded posture — the concrete facts this grade is built from. A full audit verifies each against the vendor's current documentation.
Vannus grades Deepgram against nine dimensions of trust — data sovereignty, training privacy, compliance posture, operational resilience, exit portability, and more. The heaviest criterion asks whether the tool builds its own AI or resells someone else's model; where the vendor discloses it, the grade cites the vendor's own documentation. No paid placements — scoring is walled off from affiliate revenue. See the methodology →