MCP

Community AI and agents

3D product viewer platform: upload GLB models, style scenes with lighting, camera and backgrounds, and embed them on websites.

What can you do with MCP?

  • "Upload this shoe model and give me an embed code"
  • "Make the lighting warmer and add bloom"
  • "List my models and publish the chair"
  • "Create a Look called 'Noir' with dark background and dramatic lighting"

Install

The installation steps are in the documentation below. Documentation

Documentation

From the project's README on GitHub, refreshed weekly. Links open on GitHub.

Contents
  1. vitrine MCP server
  2. Setup
  3. What you can do
  4. Tools
  5. Resources
  6. HTTP transport
  7. How it works
  8. Documentation
  9. License

vitrine MCP server

MCP server for vitrine — the 3D product viewer platform. Upload GLB models, configure scenes, publish embeds, and manage Looks from any AI agent.

Setup

Add to your MCP client config (Claude Code, Cursor, Windsurf, etc.):

{
  "mcpServers": {
    "vitrine": {
      "command": "npx",
      "args": ["@vitrine3d/mcp"]
    }
  }
}

Works immediately with no API key — anonymous uploads expire after 48 hours.

Keep your models

Say "log me in" to your AI agent. A browser window opens, you sign in (or sign up for free), and your models become permanent. No manual key copying needed.

Or set an API key manually:

{
  "mcpServers": {
    "vitrine": {
      "command": "npx",
      "args": ["@vitrine3d/mcp"],
      "env": {
        "VITRINE_API_KEY": "vt_..."
      }
    }
  }
}

Generate keys at app.vitrine3d.com → Account → API Keys.

What you can do

Ask your AI agent things like:

  • "Upload this shoe model and give me an embed code"
  • "Make the lighting warmer and add bloom"
  • "List my models and publish the chair"
  • "Create a Look called 'Noir' with dark background and dramatic lighting"

Tools

Models
ToolDescription
vitrine_list_modelsList all models with IDs, names, and publish status
vitrine_upload_modelUpload a GLB file from disk
vitrine_model_infoGet model details, thumbnail, and merged config
vitrine_delete_modelDelete a model and its storage
Configuration
ToolDescription
vitrine_get_configGet the full scene config for a model
vitrine_set_configApply a config patch (lighting, camera, background, effects)
vitrine_list_hdrisList available HDRI environment presets
vitrine_config_schemaGet the full config JSON schema
Publishing
ToolDescription
vitrine_publishPublish or unpublish a model
vitrine_get_embedGet ready-to-paste embed HTML
Looks
ToolDescription
vitrine_list_looksList your saved Looks
vitrine_apply_lookApply a Look to a model
vitrine_create_lookSave a config as a named Look
Account
ToolDescription
vitrine_account_infoCurrent plan, usage, and limits
vitrine_loginConnect your account via browser
vitrine_logoutRemove saved credentials
vitrine_feedbackSubmit a bug report or feature request

Resources

ResourceDescription
vitrine://config-schemaFull SceneConfig JSON schema
vitrine://hdri-presetsAvailable HDRI presets
vitrine://embed-templateEmbed code template

HTTP transport

Run as a hosted HTTP server instead of stdio:

npx @vitrine3d/mcp --http          # default port 3000
PORT=8080 npx @vitrine3d/mcp --http # custom port

The server exposes a Streamable HTTP endpoint at /mcp, compatible with Smithery, Glama connectors, and any MCP client that supports HTTP transport.

How it works

You -> AI Agent (Claude, Cursor, etc.)
        | MCP protocol (stdio or HTTP)
      vitrine MCP server
        | HTTPS
      api.vitrine3d.com
        |
      Supabase (database, storage, auth)

Documentation

License

MIT

Frequently asked questions

What is MCP?

3D product viewer platform: upload GLB models, style scenes with lighting, camera and backgrounds, and embed them on websites.

How do I install MCP?

Follow the installation steps in its documentation, shown on this page and in its repository.

Is MCP open source?

Yes. Its code is on GitHub under the MIT license.

Who maintains MCP?

It is published by vitrine3d. AI.tn lists it but does not maintain or endorse it.

From the community awesome-mcp-servers lists and the project's GitHub repository. Listing is not an endorsement: review a server's code and permissions before giving it access to your data.