Shared engine and interface
Local and cloud delivery modes share the classification engine and browser interface, with environment-specific behavior kept at the edge.
Reference implementation
A real engineering pattern for moving a working local application into a cloud service while preserving version control, deployment traceability, operating safeguards, and an explicit recovery path.
This is a technical reference implementation, not a client success story. MailSweep is a separate product brand. It is not a Tectori service line, and no client name, testimonial, or client result is implied here.
The challenge
MailSweep began as a single-user Windows application for mailbox triage. Its cloud variant needed to support Microsoft 365 in a hosted operating model while sharing the same core engine and user interface with the local product.
The engineering question was larger than where to run the code. The deployment needed clear version identity, controlled access to Azure, separated runtime storage, audit logging, a repeatable test path, and a way to detect drift between shared local and cloud behavior.
The implementation
Local and cloud delivery modes share the classification engine and browser interface, with environment-specific behavior kept at the edge.
The cloud service runs as a containerized application on Azure, with Azure storage services used for hosted state.
The hosted service has a deliberate Microsoft and Azure boundary instead of claiming provider support the cloud product does not implement.
Operating safeguards
Application, deployment, and infrastructure definitions are versioned together in the project repository.
The deployment pipeline uses identity federation to authenticate to Azure without storing a long-lived cloud password in the workflow.
Each production deployment points to a container image tagged with the full source commit identifier.
Automated checks detect drift in shared administrative logic between the local and cloud delivery modes.
Release notes and a running manual test plan document what changed, how to verify it, and which surface is affected.
Current state
The cloud service runs in production on Azure Container Apps. The repository records the product version, source commit, container image, deployment definition, shared code boundaries, and operator testing expectations.
This implementation does not claim a client outcome or a fabricated performance result. Its value as evidence is simpler. It shows the build standards in working code and operating records, not only in a slide deck.