SmartCompanion

Create and edit museum audioguide apps with your AI assistant: stations, tours, texts in several languages, images, audio and app design. Remote server with OAuth; a free SmartCompanion account is required.

Размещённый MCP-сервер

npx add-mcp 'https://admin.smartcompanion.app/mcp'

Устанавливается в Claude Code, Codex, Cursor и другие

Документация

SmartCompanion MCP Server

With the SmartCompanion MCP server, you manage your audioguide directly from an AI assistant such as Claude, ChatGPT, Cursor or VS Code: create stations and tours, shorten and translate texts, add images and audio, or change the design of the app. The server uses the open Model Context Protocol (MCP) and is meant for museums, castles and exhibitions that use SmartCompanion.

https://admin.smartcompanion.app/mcp

You need a SmartCompanion account. Create a free account

Connect your assistant

In every assistant, you add the server URL above and then sign in with your SmartCompanion account. Menu names may change with new versions of the assistants.

Claude (web and desktop)

  1. Open Customize › Connectors and choose “+ Add” › “Add custom connector”.
  2. Enter a name, e.g. “SmartCompanion”, and the server URL. Confirm with Continue and Add.
  3. Sign in with your SmartCompanion account and confirm the access.
  4. Enable the connector in a chat via “+” › Connectors.

Custom connectors are available on all Claude plans; the free plan allows one custom connector. In Team and Enterprise accounts, admins add the connector under Organization settings › Connectors. Details: Anthropic's guide.

Claude Code

claude mcp add --transport http smartcompanion https://admin.smartcompanion.app/mcp

Then start Claude Code, run /mcp and sign in to the “smartcompanion” server. If you have already added the connector in Claude on the web, it is available in Claude Code automatically.

ChatGPT

  1. In ChatGPT on the web, open Plugins and choose “+” › “Add custom MCP server”.
  2. Enter a name, the server URL under Connection, and OAuth as authentication.
  3. Accept the notice and choose “Create as a plugin”. Install the plugin and sign in.
  4. Select the server in a new chat with @.

Custom MCP servers are not available on every ChatGPT plan and only on the web; in company workspaces, admins decide who can add them. Details: OpenAI's guide.

Cursor

Add the server to ~/.cursor/mcp.json (all projects) or .cursor/mcp.json (one project):

{
  "mcpServers": {
    "smartcompanion": {
      "url": "https://admin.smartcompanion.app/mcp"
    }
  }
}

VS Code

Add the server to .vscode/mcp.json or to the MCP configuration of your profile:

{
  "servers": {
    "smartcompanion": {
      "type": "http",
      "url": "https://admin.smartcompanion.app/mcp"
    }
  }
}

On the first connection, your browser opens for sign-in.

Sign-in and permissions

  • Sign in with your SmartCompanion account (OAuth 2.1): you sign in in your browser and confirm the access. No API keys are needed, and your password is never shared with the assistant.
  • Automatic registration: the assistant registers itself with the server on the first connection (Dynamic Client Registration), so there is nothing to set up manually.
  • Same permissions as your account: the assistant only sees and changes the audioguides you can edit in the admin.
  • Revocable at any time: all connected assistants are listed in the admin under “AI connections”, where you can revoke access at any time.

Tools

The server provides 25 tools. The “Type” column shows whether a tool only reads, changes or deletes something; many assistants ask before changing or deleting actions.

ToolWhat it doesType
Audioguides
list-audioguidesLists the audioguides you can edit, with their slug, languages and number of stations.Read
get-audioguideShows the settings and design of an audioguide and the state of its preview, with problems that would block publishing.Read
Stations
list-stationsLists the stations with title, number of images and audios, and missing translations.Read
get-stationShows a station with all translations, its location, images and audios.Read
create-stationCreates a new station.Change
update-stationChanges a station's number, texts (some or all languages) or location.Change
delete-stationDeletes a station with its audios and removes it from all tours.Delete
add-station-imageAdds an image to a station (JPEG, PNG, WebP or AVIF, max. 20 MB).Change
remove-station-imageRemoves an image from a station.Delete
add-station-audioAdds an audio in one language to a station (MP3, M4A, AAC, OGG or WAV, max. 200 MB).Change
update-station-audioChanges the language, title or position of an audio.Change
remove-station-audioRemoves an audio from a station.Delete
Tours
list-toursLists the tours with their stations in visiting order.Read
get-tourShows a tour with all translations, its image and its stations.Read
create-tourCreates a tour with its stations in visiting order.Change
update-tourChanges a tour's number, texts, default tour setting or stations.Change
delete-tourDeletes a tour; its stations are kept.Delete
set-tour-imageSets or removes the image of a tour.Change
App texts
list-textsLists the texts of the app's interface (buttons, menus, messages) with their translations.Read
update-textsChanges texts of the app's interface.Change
Branding and settings
update-brandingChanges the app's name, short name, description and colours.Change
set-branding-imageSets or removes the logo, the dark mode logo or the app icon.Change
update-app-settingsChanges the content languages, the URL of the share button and the app's PIN codes.Change
Uploads
create-upload-linkCreates an upload link for images and audio files, valid for one hour.Change
list-uploadsLists the uploaded files with their upload ids.Read

The assistant passes images, audio files and logos as a public URL or through an upload link: with create-upload-link it creates a link where you upload files in your browser. This keeps the quality and metadata of your files.

Drafts and publishing

New accounts start with a sample audioguide, so you can try things out with real content right away. All changes made via MCP go to the draft of your audioguide: they appear within seconds in the unpublished preview in the admin, where you can check them. Publishing is unlocked by SmartCompanion after a short review; after that, you publish yourself in the admin. Publishing is not possible via MCP, but get-audioguide shows whether publishing has been unlocked yet.

Examples

  • "Show me all tours of my audioguide."
  • "Create a tour with eight stations for the new temporary exhibition. The texts are in the attached document."
  • "Shorten all station texts of the 'Knights' Hall' tour to at most 120 words and keep the tone."
  • "Add the English translations to all stations where they are still missing."
  • "Replace the cover image of the family tour with the photo I just uploaded."
  • "Change the primary colour of the app to match our new corporate design."

Data protection

Through the MCP server, SmartCompanion processes your account data for sign-in, the access tokens of the connected assistant, and the content of your audioguide that the assistant reads or changes. Your AI provider (e.g. Anthropic or OpenAI) processes this content under its own terms. No personal data of your app's visitors is involved. More in our privacy policy.

Troubleshooting and support

  • The URL shows an error (405) in the browser: that's expected – the server is only meant for AI assistants, not for browsers.
  • Sign-in fails: you need a SmartCompanion account first. Also check that the URL is exactly https://admin.smartcompanion.app/mcp.
  • The assistant doesn't see an audioguide: your account has no audioguide yet or no permission for it.
  • New tools don't show up: start a new chat or reconnect the server.

Questions? Email us at hello@smartcompanion.app.