Middleware and orchestration

When to use: Your integration sits between Speckle and enterprise systems (PLM, data lakes, scheduling, proprietary APIs) without owning a host-application UI. Typical stack:

GraphQL API and REST where documented
specklepy or @speckle/objectsender / @speckle/objectloader2 depending on runtime
Idempotent, version-aware sync patterns

Docs: Key concepts · Updating syncs · Webhooks

Design notes

Treat versions as immutable inputs; design jobs to be replayable from a known version id.
Separate read paths (pull from Speckle) from write paths (push new versions) so failures are easier to reason about.
Prefer explicit mapping tables for external system ids ↔ Speckle applicationId values.

Data & build order
Desktop host · Cloud & automation · Hybrid
Resources

Last modified on May 19, 2026

Cloud & Automation Hybrid

​Design notes

​Related

Design notes

Related