Hybrid connectivity

Combine desktop publish, cloud processing, and optional web consumption in a multi-tier Speckle integration.

When to use: A desktop connector or agent publishes data; cloud jobs or other services consume the same project, model, and version identifiers—often alongside a separate web product for viewing or workflows.

Typical boundaries

1. Extract and convert in the host — see Desktop host
2. Serialize and transport via SDK — create a Speckle version
3. Consume downstream — cloud jobs (Cloud & automation), middleware (Middleware), or UI in Applications (Custom apps, Viewer)

The browser/UI layer is usually Applications documentation, not a connectivity architecture on its own. Keep host and cloud boundaries explicit so teams do not duplicate viewer logic inside connectors.

Stacks by tier

Tier	Typical technology	Documentation
Host	.NET connector, speckle-sharp-sdk	Desktop host
Transport	GraphQL, specklepy operations	GraphQL API
Cloud processing	specklepy, webhooks	Cloud & automation
Web UI	Viewer, custom apps	Custom apps

Related

• Data & build order
• Resources