Applies to every architecture in this section. For full detail see Data schema. Before the steps below, decide publish only or publish and load for your host (receive-only is rare—mainly consumption apps like Power BI). See Publish & load for integration depth.Documentation Index
Fetch the complete documentation index at: https://docs.speckle.systems/llms.txt
Use this file to discover all available pages before exploring further.
Projects, models, and versions
- Project — groups work for a team or deliverable
- Model — logical dataset (often per discipline or file)
- Version — immutable snapshot after a send; design send, receive, and automation around versions
Extraction, conversion, serialization, transport
Speckle exchanges objects in a directed acyclic graph, not opaque files:- Extraction — read native entities or files
- Conversion — map to Speckle types and metadata
- Serialization — build the graph (collections, proxies)
- Transport — authenticate, send, create a version
Semantic data, geometry, and display values
- Semantic data — Object schema
- Geometry — Geometry schema
- Display values — packaging for the Viewer; see Display values
displayValue (or equivalent) so viewers and downstream tools stay aligned.
Suggested build order
Skip steps that do not apply (for example receive-only integrations).Authentication
PAT, OAuth2, or app credentials — Authentication.
Minimal object send
Create a project and model, send a trivial graph, confirm a version in the web app. Python quickstart or Object Sender.
Geometry extraction
Convert primitives your consumers need — Geometry schema.
Incremental updates
Define change sets (selection, delta, full resend) and version relationships for sync.
Performance
Batch and deduplicate — Optimise complexity.
Receive and update
Bidirectional flows: study receive paths in speckle-sharp-connectors.
Deployment
Installers, logging, updates — Manual installation for .NET connector patterns.
Architectures
Publish & load
Publish/load direction and integration depth.
Desktop host
Host add-ins and native desktop agents.
Cloud & automation
ETL, orchestration, and version-driven jobs.
Middleware
Enterprise orchestration between Speckle and other systems.
Hybrid
Multi-tier publish and consume patterns.