Infrastructure & platform teams
Deploy within your operating boundary.
Choose where capabilities run and which services they can reach. Separate desktop authoring, team services and mobile participation, then review credentials, data routes and recovery for each environment.
Launch specification. Application source, downloads and versioned compatibility evidence publish with releases.
Operating workflow
Self-hosted deployment, from input to evidence.
A deployment plan tied to package versions, operating roles, credential scope, approved provider routes and a tested recovery procedure.
- 01
Choose the runtime role
Map authoring and review to desktop workspaces; map persistent operations to team services. Docker and Compose provide repeatable service packaging. Mobile roles remain separately bounded.
Review the specification - 02
Review the connection boundary
Check the adapter version, supported operations, identity and credential scope. Record where data is processed; self-hosting a control plane does not make every provider local.
Review the specification - 03
Prepare rollout and recovery
Keep configuration, secrets, state and release artifacts distinct. Define health checks, rollback conditions, receipt retention and reconciliation ownership before deployment.
Review the specification
Different systems. Explicit roles.
macOSDesktop
WindowsDesktop
LinuxDesktop / server
DockerService packagingiOS / iPadOSCompanion / SDK
AndroidCompanion / edgeOperator experience
Inspect the configuration behind the action.
Use the interface preview to orient the evaluation. Review the linked specifications for boundaries, responsibilities and evidence.
Evaluation checklist
Define a bounded first workload.
Work through these questions, then continue to the technical material for your operating role.
Desktop or persistent service?
Use the operating role to select a deployment form. A desktop authoring workflow and a shared service have different uptime, identity and state requirements. Release documentation identifies supported packages and lifecycle requirements.
What remains inside the boundary?
List model endpoints, connector destinations, secret stores and operational records. Document approved egress and processing locations. Local infrastructure can still send data to a remote provider if the workflow permits it.
What changes on mobile?
iOS, iPadOS and Android participate through companion workflows, policy SDKs and explicitly bounded edge work. Review platform lifecycle, background execution and permission limits separately from desktop and server capabilities.
