The delivery process
Product Files
Bring your product and the files you have. We define the preparation, build the guided experience, and publish the approved system where your customers and teams need it.
Start with what you have.
File readiness changes the preparation work—not automatically the subscription tier. The system also depends on the product count, interactions, deployment, and support required.
Existing GLB or 3D model
We review geometry, materials, structure, and performance. A file that looks complete may still need optimization or separated parts before it can demonstrate movement.
CAD or engineering files
Conversion and web preparation are scoped after file review. Engineering geometry may need simplification, material work, and interaction-ready structure.
A static model that needs motion
Share movement references and the parts that need to move. Separation, articulation, or reconstruction may be required; a GLB extension alone does not confirm readiness.
No usable 3D files yet
Start with product images, dimensions, specifications, and approved references. Model creation is assessed and quoted separately before work begins.
Review the product and agree the scope.
Share the product, available files, intended audience, and what people need to understand. We identify preparation needs, interactions, deployment requirements, and the work to be approved before production.
What this produces
An agreed scope separating asset preparation, viewer setup, and ongoing operations.
Prepare the product foundation.
Approved source material is prepared for the web: geometry, materials, product structure, and the states needed for the experience. Conversion, model creation, and articulation are scoped according to the starting assets.
What this produces
A viewer-ready product foundation built around the approved requirements.
Build and review the experience.
We shape the product into guided views for inspection, movement, configurations, or explanation. Your team reviews the product representation and content before the experience is approved for release.
What this produces
An approved experience, with the agreed interactions and product information.
Deploy and keep it useful.
Publish through supported website embeds, direct links, or other agreed destinations. Continuing hosting, updates, quality checks, and support follow the operating scope you approve.
What this produces
A hosted product experience with a defined path for updates and support.
Know what is included before you commit.
Your written scope confirms the deliverables, asset preparation, supported destinations, update allowances, and support commitments. Extra products, new interactions, or work outside that scope require separate approval.
The pricing estimator is planning guidance, not checkout or automatic file acceptance. Submitting a scope request does not take payment.
How we use AI
Real product models. Not AI-generated.
We build models of real products from product files, documented dimensions, and approved references—not AI-generated geometry. We use AI throughout production to assist preparation, presentation, and quality checks, including generating environments for the models to sit in. AI supports the work; it does not invent the product.
The real product comes first.
REALGINE builds and prepares product models from supplied files, documented dimensions, and approved references. We do not use AI to generate product models or invent geometry, internal parts, or mechanical behaviour.
AI assists with asset cleanup, component identification, and extracting information from documentation. Those outputs are checked against the sources; cleanup must preserve the product’s documented geometry and dimensions.
Human approval comes before publication.
Illustrative workflow · Not live system activityAI-assisted checks support human review; they do not establish dimensional accuracy, validate physical behaviour, or provide engineering, safety, or regulatory certification.
Missing dimensions, specifications, or movement information must be resolved with approved sources—not filled in by AI. Product facts remain grounded in the real product.