Continuity

Backup & disaster recovery

This page is maintained by ScribeMDPro to explain how clinical data is protected, how quickly service is restored after a disruption, and how the platform scales internationally.

What is and is not stored

Audio recordings
Not retained. Consultation audio is processed to produce a transcript and is not kept as a stored recording, so there is no audio archive to back up or breach.
Transcripts and notes
Stored against the owning clinician or clinic workspace and protected by database-enforced access rules.
Account data
Email, name, specialty, preferred language and subscription state.
Operational metadata
Aggregate usage counters and export records used for billing accuracy and workspace accountability.
Support and analytics
Product analytics carry page and interaction metadata only — never transcripts, note content or patient identifiers.

Backup

Managed, encrypted backups of the production database.

Frequency

Automated daily backups of the production database, plus continuous transaction logging that enables point-in-time recovery between daily snapshots.

Available
Encryption

Backups are encrypted at rest by the managed database platform and are not accessible from the application runtime.

Available
Restore testing

Scheduled restore drills that verify a backup can be recovered into a clean environment, with the result recorded.

Planned — not yet active

Recovery objectives

Targets we design and operate against. Contractual objectives for hospital deployments are agreed in writing during procurement.

RPO — Recovery Point Objective
Target of under 24 hours from daily snapshots, reduced toward minutes where point-in-time recovery is available.
RTO — Recovery Time Objective
Target of under 4 hours to restore core consultation, transcription and note functionality after a declared incident.
Degraded mode
If generation services are unavailable, previously created notes remain readable and exportable — clinicians never lose access to documentation they already completed.
Contractual SLAs
Enterprise and hospital agreements can set specific availability and recovery commitments; the targets above are operational, not a public guarantee.

Data retention & deletion

Active accounts
Transcripts and notes are retained while the account or workspace is active, so clinicians can revisit past documentation.
Deletion requests
Clinicians can delete their own consultations and patient records; workspace owners and admins can remove records within their clinic.
Backup lag
Deleted data may persist in encrypted backups until those backups age out of the retention window, after which it is unrecoverable.
Account closure
Contact us to close an account and remove its data. Aggregate, non-identifying usage counters may be retained for billing and statutory record-keeping.

Incident response

  1. 1

    Detect

    Automated error and availability monitoring, plus reports through the support channel.

  2. 2

    Triage

    Severity assessed against clinical impact first: can clinicians record, generate and export?

  3. 3

    Contain

    Restrict the affected surface, revoke compromised credentials, and roll back or disable the failing change.

  4. 4

    Recover

    Restore service from backup or a known-good deployment and verify with end-to-end checks.

  5. 5

    Notify

    Affected workspace owners are informed with what happened, what data was involved and what to do — without exposing patient content in the notification.

  6. 6

    Review

    Written post-incident review with corrective actions tracked to completion.

Report a suspected security issue or outage to scribemdpro@gmail.com. Please do not include patient-identifying information in the report.

Global scalability

How the platform is architected to serve clinicians across regions.

Edge delivery

The application is served from a globally distributed edge network, so pages and assets load close to the clinician regardless of country.

Available
Stateless compute

Server work runs on stateless, horizontally scalable workers with no machine-local state, so capacity grows with demand.

Available
Timezone & currency

Dates render in the viewer's locale and pricing is presented per region; per-workspace timezone and currency preferences are planned.

Planned — not yet active
Multi-region data residency

Region-pinned data storage for jurisdictions that require it, offered per enterprise deployment.

Planned — not yet active
Regional read replicas

Lower-latency reads for large hospital networks operating far from the primary region.

Planned — not yet active
Interface localisation

Dictation already spans many languages; a translation framework for the interface itself is prepared, and languages are enabled only once fully translated.

Planned — not yet active

Related

Continuity questions before you deploy?

We will walk your IT team through the architecture in detail.