Library › About ·

Real-world geometry for AI assistants.

Xem Studio is a searchable geometry library and a free MCP server for AI assistants. Connect it to a compatible AI tool so it can look up physical object dimensions, mating interfaces, and available STEP/STL CAD references while designing parts that fit real things.

Connect Xem to your AIBrowse the geometry library

Free MCP connection · No Xem account or API key required · Compatible-client support varies

Why Xem exists

An AI can design a holder, adapter, replacement knob or mounting bracket, but it still needs a trustworthy reference for the object it must fit. Guessing from photos or overall dimensions can lead to failed prints. Xem gives people and AI agents a place to find real-world geometry, fit details, and source information instead of starting each design from scratch.

The site can be searched normally, but the intended repeat workflow is to connect Xem Studio to your AI assistant using the Model Context Protocol (MCP). Your assistant can then query the library from a supported chat or CAD workflow when it needs a reference.

Connect once, use Xem in your AI workflow

  1. Connect. Claude: add Xem from the Claude directory in one click. ChatGPT, Cursor or VS Code: add https://xemstudio.com/mcp as a remote MCP server. Setup by platform.
  2. Ask for a real object. Have your assistant find the dimensions of a device, a hole pattern, a standard fitting, or a mating surface before it creates a new part.
  3. Design using the reference. Use the available dimension sheets, fit records and licensed STEP/STL files. Where a required object is missing, your assistant can request a new geometry entry.

A connection makes the tools available to supported clients; it does not mean an AI will always use them automatically. Check the source and stated accuracy of each geometry record before relying on a critical fit.

Where the geometry comes from

The distinction matters: a nominal dimension, a manufacturer-sourced dimension and a measured scan are different kinds of evidence. Xem aims to make those differences clear so an AI agent can decide what reference is appropriate.

For developers and AI agents

The remote MCP connection guide covers supported assistants. The HTTP geometry API and OpenAPI specification provide machine-readable access to catalog metadata, dimensions, availability and licenses. Object pages also offer Markdown representations, and the AI-readable index helps clients that do not support MCP.

What's live: the free catalog, dimension sheets, available hardware CAD files, public reference links, MCP tools and HTTP API. What's planned: a broader library of scanned physical objects, premium mesh/STEP downloads and agent payments via x402. See current and planned pricing.

Frequently asked questions

Is Xem Studio a CAD download website or an AI connection?

Both. You can browse the catalog in a browser, but Xem's MCP server lets compatible AI assistants look up the same geometry within their own workflows. Connect your AI here.

Can Claude or ChatGPT use Xem Studio?

Yes. In Claude it is one click: add Xem Studio from the Claude directory. In ChatGPT it is a custom connector on plans that support them. See the platform-specific setup instructions.

Are all Xem dimensions independently 3D scanned?

No. The library includes standards-based geometry, manufacturer or public references, and measured information where available. Check each object's provenance, status, fit notes and tolerances before manufacturing a part.

Can an AI buy a file?

Yes. Agents pay per file with x402 (USDC on Base, no account): the server answers 402 with the price, the agent pays and retries. Live on mainnet; the x402 page lists what is for sale. Scanned meshes and STEP solids will be sold the same way when they publish.

Can I request an object that isn't listed?

Yes. Submit the object and the fit you need, or ask a connected AI to use Xem's request tool. Requests help determine what gets added next.

Who builds Xem Studio?

Xem Studio is operated from Mesa, Arizona, by Long Earth Society. Its engineering work with Stirling and thermoacoustic engines, standard hardware and physical interfaces informs the library's approach to reusable reference geometry.

Founded by Travis Seamon, MS, CSP, NRP, creator of Long Earth Society. This site was built to assist AI with designing around standard equipment and building materials.