Cloud Speech Engines
Use cloud speech-to-text providers intentionally and understand what leaves your device.
MachinesFluent can also use cloud speech-to-text providers. Choose a cloud engine when you prefer that provider's transcription behavior, language support, speed, or accuracy for a specific job.
The Important Privacy Difference
When you choose a cloud speech engine, speech recognition is no longer fully local. Audio or audio-derived data must be sent to the selected provider so the provider can return a transcript.
Provider-side storage, retention, training, logging, billing, and compliance behavior are governed by that provider's own terms and settings.
Supported Provider Set
The current product context lists these cloud STT providers:
- AssemblyAI
- Deepgram
- Cartesia
- Gladia
- Gradium
- Groq
Availability can vary by build, account setup, provider availability, and license tier.
When Cloud STT Makes Sense
Cloud STT may be useful when:
- a provider handles your language or accent better
- you need a provider-specific model
- local hardware is too slow for your workflow
- you are intentionally using a paid external speech provider
- privacy requirements are lower than accuracy, latency, or convenience requirements
Before Sending Sensitive Audio
Check:
- Which provider is selected.
- Whether the provider requires an API key or account.
- The provider's current data handling terms.
- Whether the audio contains private, legal, medical, customer, or company-confidential information.
- Whether your organization allows that provider.
Do not assume cloud privacy from local docs
MachinesFluent can offer local dictation paths, but cloud engines are external provider workflows. Treat cloud STT as a deliberate data-routing choice.