Deepgram
Speech recognition and speech generation APIs, including real-time transcription used inside voice agents. A component, not a managed service.
- Supplier layer
- Speech/model infrastructure
- Workflow fit
- Enterprise
- Deployment responsibility
- Build (assemble)
- Directions
- inbound, outbound
- Pricing basis
- Usage-based per audio minute; confirm current pricing.
- Integration evidence
- Component layer: no business-system integrations are implied.
- Transfer evidence
- Not applicable at this layer.
- Data and retention documentation
- Not yet reviewed by Trace Margin. Ask the supplier for documentation.
- Support
- Plan-dependent.
- Portability
- Component swap is usually feasible; confirm contract terms.
- Verification status
- Category-level; not tested by Trace Margin