MCP App Store
by KeenEthics
Rayon icon

Rayon MCP Server for ChatGPT

By RAYON

Overview

Develop interior and architectural designs directly in chat. Refine floor plans, try different furniture layouts, visualize design ideas, and prepare schedules and client presentations using your connected Rayon projects.

How to connect

  1. ChatGPT: Open the Rayon listing in the ChatGPT app directory and choose Connect. ChatGPT listing Rayon exposes 38 tools here.

Tools

annotation

ChatGPT
Rayon annotation operations. Choose a subcommand with operation and supply its matching input. In run_typescript, the same functions live under api.annotation and accept the same input. The subcommand schemas provide input documentation.

assets

ChatGPT
Rayon assets operations. Choose a subcommand with operation and supply its matching input. In run_typescript, the same functions live under api.assets and accept the same input. The subcommand schemas provide input documentation.

block

ChatGPT
Rayon block operations. Choose a subcommand with operation and supply its matching input. In run_typescript, the same functions live under api.block and accept the same input. The subcommand schemas provide input documentation.

camera

ChatGPT
Rayon camera operations. Choose a subcommand with operation and supply its matching input. In run_typescript, the same functions live under api.camera and accept the same input. The subcommand schemas provide input documentation.

canvas

ChatGPT
Rayon canvas operations. Choose a subcommand with operation and supply its matching input. In run_typescript, the same functions live under api.canvas and accept the same input. The subcommand schemas provide input documentation.

completeDrawingUpload

ChatGPT
Validate and register a DWG uploaded using prepareDrawingUpload. Returns a fileId for importDrawing. Retrying completion is safe.

createModel

ChatGPT
Create an empty Rayon model with the specified name in your personal workspace. Returns its ID for use with model tools.

dataTables

ChatGPT
Rayon dataTables operations. Choose a subcommand with operation and supply its matching input. In run_typescript, the same functions live under api.dataTables and accept the same input. The subcommand schemas provide input documentation.

datum

ChatGPT
Rayon datum operations. Choose a subcommand with operation and supply its matching input. In run_typescript, the same functions live under api.datum and accept the same input. The subcommand schemas provide input documentation.

entity

ChatGPT
Rayon entity operations. Choose a subcommand with operation and supply its matching input. In run_typescript, the same functions live under api.entity and accept the same input. The subcommand schemas provide input documentation.

getModel

ChatGPT
Get metadata for one accessible Rayon model.

getModelFile

ChatGPT
Sign a short-lived download URL for a model the user may read. Called by the Rayon viewer.

importDrawing

ChatGPT
Create a new private Rayon model from a DWG. Pass either a host-provided attachment in file or an uploaded Rayon fileId, not both. Attachment upload is handled internally. For clients without file inputs, use prepareDrawingUpload and completeDrawingUpload first. Returns modelId, status and fileId after starting conversion; use waitForDrawing until Ready or Failed. Defaults: import paper spaces, one canvas per page, remove empty layers and unused blocks, preserve hidden layers, center at origin, no simplification. Source DWG is retained. Conversion may take several minutes. Inspect dimensions and unit text to establish scale after importing. For insertion into an EXISTING model, select the destination model_id, call start_import with a stored fileId, call wait_for_job until Ready, then use importExport.importModel with the conversion result's modelId, an explicit sourceCanvas and a destination layer.

import_status

ChatGPT
Returns only the conversion status: Converting, Ready or Failed. Reads status immediately without waiting or downloading model data. Use wait_for_job for asynchronous waiting.

inspect_scene

ChatGPT
Inspect the live Rayon scene as budgeted semantic XML with hierarchy, styles, world bounds and labeled transforms. Space-3D canvases additionally include solid kinds, 3D world bounds and local 3D transforms. Set include_visuals: true for an image and labeled minimap. Canvas inspections give an overview; handle, cluster and region inspections provide larger images for reading individual drawings. Use target: {kind: 'region', canvas: canvasHandle, bounds: {min: {x, y}, max: {x, y}}} to inspect a specific 2D drawing crop in source-world millimetres, without creating a page. Omitted-child comments provide follow-up targets and continuation cursors. The initial prompt already contains a whole-model overview; focus subsequent inspection on useful subparts or changed content.

jobs

ChatGPT
Rayon jobs operations. Choose a subcommand with operation and supply its matching input. In run_typescript, the same functions live under api.jobs and accept the same input. The subcommand schemas provide input documentation.

layer

ChatGPT
Rayon layer operations. Choose a subcommand with operation and supply its matching input. In run_typescript, the same functions live under api.layer and accept the same input. The subcommand schemas provide input documentation.

listModels

ChatGPT
List accessible Rayon models. Use sort_by createdAt and sort_order desc for recently created models, or updatedAt desc for recently modified models.

listOrganizations

ChatGPT
List Rayon organizations visible to the authenticated user.

listWorkspaces

ChatGPT
List Rayon workspaces visible to the authenticated user.

list_skills

ChatGPT
List the available skills: each skill's SKILL.md URI, when to use it, and the URIs of its files.

parametric

ChatGPT
Rayon parametric operations. Choose a subcommand with operation and supply its matching input. In run_typescript, the same functions live under api.parametric and accept the same input. The subcommand schemas provide input documentation.

prepareDrawingUpload

ChatGPT
Prepare a DWG upload. Returns fileId, url and fields. From your file-capable sandbox, POST multipart/form-data to url with every returned field and the original binary file as the last field named file. Then call completeDrawingUpload. Requires network access to the upload URL; otherwise upload through Rayon.

presentation

ChatGPT
Rayon presentation operations. Choose a subcommand with operation and supply its matching input. In run_typescript, the same functions live under api.presentation and accept the same input. The subcommand schemas provide input documentation.

property

ChatGPT
Rayon property operations. Choose a subcommand with operation and supply its matching input. In run_typescript, the same functions live under api.property and accept the same input. The subcommand schemas provide input documentation.

read_doc

ChatGPT
Read a whole Rayon documentation page or interior design guide article as Markdown, by a URL from search_docs.

read_skill

ChatGPT
Return the content of a skill file by its URI from list_skills.

report_feedback

ChatGPT
Report an observed Rayon MCP issue or user-requested feedback to the Rayon team. Include model_id when the issue concerns a model. Call at most once per distinct issue; at most 20 reports per user per hour. Do not copy model data, credentials, or other personal information into the report.

run_typescript

ChatGPT
Execute a focused TypeScript script against the loaded Rayon model. Use the API operation documentation provided in the current context. Scripts use module namespaces such as api.entity and api.presentation. Write awaited API calls as top-level statements; each call applies immediately in order. Coordinates and distances use millimeters. Scripts are typechecked before running (strict, noUncheckedIndexedAccess, exactOptionalPropertyTypes); type errors apply no model changes. - arr[i] may be undefined: check it, or write arr[i]! when it must exist. - Omit optional fields instead of setting them to undefined. The sandbox rejects imports and limits heap usage, execution time, and operation count. Successful scripts return only {logs: [...]}, with an empty array when nothing was logged. Use console.log to return the handles, counts, or selected fields you need; objects are printed as JSON and output beyond about 20 KB is truncated. Script errors include an API execution record for recovery. Pass canvas, layer, style, and entity handles explicitly; operations act only on the handles they receive. The script is interpreted as a toplevel Typescript statement. Do not return from it, as it would be illegal syntax.

screenshot

ChatGPT
Capture a saved camera, entity, canvas, cluster, or every page of a canvas in the loaded Rayon model. 2D images include a labeled SVG minimap; 3D images render solids and meshes. The JSON metadata identifies the canvas and image size; automatically framed targets also include world-space bounds and discoverable clusters.

search_docs

ChatGPT
Search Rayon's product documentation and interior design guide, including the Lexicon. Set source to design_guide for design terms, dimensions and materials, or rayon_docs for how to use Rayon. Returns up to 10 results with links; use read_doc for a full page.

showModel

ChatGPT
Show the user a Rayon model's live canvas in an interactive read-only viewer. Without a view it opens the model's active canvas, fitted to its content.

showPresentation

ChatGPT
Show the user a presentation of a Rayon model in an interactive viewer. Pass presentation_id to open a given presentation or revision of that model; without it, the model's all-pages presentation opens.

solid

ChatGPT
Rayon solid operations. Choose a subcommand with operation and supply its matching input. In run_typescript, the same functions live under api.solid and accept the same input. The subcommand schemas provide input documentation.

start_import

ChatGPT
Starts DWG conversion for insertion into the CURRENT model. The returned modelId identifies a temporary conversion result, not the user's destination. Read conversion status with import_status; use the separate wait_for_job tool for asynchronous waiting. For insertion without layouts, set importPaperSpaces to false. Once Ready, call api.importExport.importFirstModelCanvas with modelId and a destination layer to insert only the first model canvas. Do not report import complete until insertion succeeds. Check the imported drawing's scale against unit text and labelled dimensions. Conversion continues if the conversation ends.

style

ChatGPT
Rayon style operations. Choose a subcommand with operation and supply its matching input. In run_typescript, the same functions live under api.style and accept the same input. The subcommand schemas provide input documentation.

waitForDrawing

ChatGPT
Read conversion status, waiting up to timeoutSeconds (0–60, default 60). Returns only the conversion status enum. If Converting, call again in a later tool call. Ready means model tools can load the model; Failed means conversion failed. A wait timeout leaves conversion running.

wait_for_job

ChatGPT
Waits asynchronously without acquiring the model's engine. For AI generation, pass { jobId } and receive the same status, result and image output fields as api.jobs.get. For DWG conversion, pass { modelId } from start_import or importDrawing and receive Converting, Ready or Failed. Provide exactly one ID. Optional timeoutSeconds is 0–60; defaults to 30 for AI jobs and 60 for drawings. Zero reads status once. If pending, running or Converting, call wait_for_job again with the same ID. Handle completion or failure before ending your turn. Cancellation stops waiting and leaves the backend task running. Call this separate tool after run_typescript submits the task.

Capabilities

WritesInteractive

Example Prompts

Click any prompt to copy it.

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/rayon"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/rayon)