Released: Apr 7, 2026
Prepared by: Operations with assistance from Engineering
Effective: Apr 7, 2026
Platform reliability, performance, and scalability
Improved handling of long‑running and continuous monitoring sessions so that audio processing remains reliable even when backend components are restarted or rotated.
Tuned infrastructure (including workflow and temporal services) with updated memory allocation and supporting node pools, improving stability under higher load.
Investigated and addressed model latency patterns using new observability and dashboard insights, helping keep scoring throughput predictable.
Resolved an issue where assessments could fail if workflow pods restarted mid‑run, reducing the chance of incomplete or failed assessments.
Data handling, reporting, and analytics
Advanced plans for a dedicated data warehouse and reporting stack to better centralize, analyze, and share assessment and usage data.
Tightened security posture around analytics services, including reducing unnecessary public exposure of analytics workspaces.
Continuous monitoring and ambient use cases
- Enhanced configuration options for continuous monitoring so that customers can choose between diarized “speaker” processing or unified audio processing, aligning better with their specific use cases.
Security, compliance, and observability
Addressed vulnerabilities and required platform updates, including OS and framework upgrades (for example, Windows and .NET runtime updates) to keep the environment aligned with current security best practices.
Implemented key and certificate lifecycle improvements for critical services, reducing operational risk from expiring credentials.
Improved security for analytics and data services by tightening network exposure and authentication paths.
Continued quarterly access reviews and related security hygiene work to ensure that infrastructure and data services only grant access where needed.
Voice AI models and scoring
- Enabled configuration and testing scenarios for augmented and specialized wellness models across internal demos and evaluation flows, preparing them for broader use.
Samsung watch and mobile experience
Refined the Samsung mobile and watch demo experience to better showcase updated wellness models and app flows.
Updated Samsung Health SDK configurations and approvals so that health‑related telemetry (such as steps and sleep) continues to flow correctly into the mobile experience.
Polished the Android app icon so that it uses a modern, full‑circle icon on compatible launchers.
Tools, integrations, and customer support readiness
Improved internal monitoring and alerting tooling (for example, OpsGenie/Sentry integrations) to ensure that alerts and incident notifications are reliable.
Added test tooling for webhook integrations to simplify and harden integration with external systems.
Continued tightening permissions and access patterns across collaboration, analytics, and infrastructure tools so teams have the access they need without over‑provisioning.
Deployment and Access:
Always keep your application up to date by enabling automatic updates. We will take care of the rest.
Contact Information:
For support or additional details, please contact:
Email: support@canaryspeech.com