Connect a 3DScribe scene to an AI assistant, then create your first one. One URL, no SDK, nothing to install.

The authoring platform behind every scene: upload CAD, structure the assembly, enrich each part, author tours and motion, then publish.

What the connector exposes once a scene is published: the tools an assistant can call, what each returns, and how answers stay tied to the model.

What an agent can do inside a published scene: select and isolate parts, play motion, run guided tours, and answer from the geometry itself.

Brand the viewer, embed it in your own pages, and decide which controls and capabilities your audience gets.

Who can open a scene, what a viewer can reach, and why your source CAD never leaves your control.

Organisations, seats and roles: inviting your team, granting access per scene, and keeping published work under one account.

Plans, usage and invoicing: what counts toward your quota, how scenes and seats are billed, and how to change plan.

What people actually do inside a scene: which components they open, what they ask, and where they hesitate.

Every release, what changed, and whether you need to do anything about it.