Spatial 3D APIs: A Practical Guide to Scenes, Spaces, and Data
Build clearer 3D data contracts with named coordinate spaces, durable object identities, and a testable scene-loading workflow.
Read the field note
YOUR DEVELOPER LEARNING PATH
Start building spatial
Choose the next decision you need to make—not the longest list of tools. These four learning paths connect the ten field notes into practical sequences for 3D foundations, immersive interaction, scene understanding, and asset delivery.
01 / Build a 3D foundation
Define coordinate conventions and scene identity, then prepare a model package you can reproduce. Your milestone is one correctly sized, correctly oriented object with one useful interaction.
glTF vs GLB: Prepare 3D Models for Reliable Spatial Delivery
Check packaging, scale, pivots, materials, dependencies, and the first useful interaction before publishing a 3D model.
Read the field note02 / Choose an immersive route
For browser VR, finish the session lifecycle before expanding the scene. For AR, move deliberately from candidate geometry to placement and attachment. Review accessibility as part of the task, not as a final decoration.
Build a WebXR VR Starter That Completes the Whole Journey
Plan browser support, deliberate session entry, clear interactions, recovery, and a useful non-immersive experience.
Read the field note
OpenXR vs WebXR: Choose the Right Layer for Your Spatial App
Compare distribution, required features, rendering choices, and maintenance before choosing an XR application architecture.
Read the field note
AR Hit Testing: From a Screen Tap to Stable Object Placement
Turn geometric candidates into an understandable placement flow with previews, confirmation, adjustment, and recovery.
Read the field note
Spatial Anchors: Keep AR Content Attached to the Right Place
Separate object identity from tracking resources and design reliable attachment, repositioning, and cleanup behavior.
Read the field note
Accessible XR: Design Beyond the Default Headset User
Make spatial tasks work through more than one input, viewpoint, or sensory channel, from session entry to exit.
Read the field note03 / Add scene intelligence
Keep geometric evidence separate from semantic predictions. Define the decision, the confidence policy, and the behavior when inputs are missing or in disagreement. Your milestone is a bounded feature that can abstain usefully.
Depth and Occlusion: Make Virtual Objects Belong in the Real World
Use depth evidence thoughtfully, test visual boundaries, and preserve useful interactions when information is missing.
Read the field note
Spatial AI APIs: Scene Understanding Without False Certainty
Design scene-understanding features that preserve context, confidence, and the distinction between inference and fact.
Read the field note04 / Make delivery repeatable
Connect model approval to a real release process. Identify assets, verify dependencies, preserve task-critical parts across quality variants, and retain a coherent revision you can roll back.
Spatial Game Assets: Build a Pipeline That Survives Real Devices
Create a repeatable content handoff with asset budgets, import checks, quality variants, release manifests, and rollback.
Read the field noteglTF vs GLB: Prepare 3D Models for Reliable Spatial Delivery
Check packaging, scale, pivots, materials, dependencies, and the first useful interaction before publishing a 3D model.
Read the field noteReady to go spatial?
Start with a clear foundation. Then build the experience your audience actually needs.
Explore the developer guides