Skip to main content

Connect AI tools to the Blackboard MCP Server

You can use the Blackboard® MCP Server from any AI assistant that supports remote MCP servers with OAuth sign-in. This topic has a short tutorial for each common AI tool. Each one adds the same server address and ends with the same sign-in and the same test.

Before you begin

  • Your Blackboard LMS administrator must turn on the Blackboard MCP Server first. If you can't find your institution when you sign in, ask them.

  • Use the AI account your institution provides. Business, enterprise and education plans keep your data out of model training by default. Personal plans may not.

  • On business, enterprise and education plans, your AI tool's administrator may need to add the server for everyone before you can connect. The tutorials below show both steps.

The server address is the same for everyone:

https://mcp.api.blackboard.com

You don't need a client ID, secret or API key. The server registers your AI tool automatically when it connects.

Sign in the first time

Every tool uses the same sign-in. The first time you use the Blackboard MCP Server, or when you select Connect:

  1. A browser window opens. Choose your institution from the list.

  2. Sign in on your usual Blackboard LMS or single sign-on page.

  3. Review the request and approve it.

  4. Return to your AI tool. You're connected.

Your password stays with Blackboard. Your AI tool receives only a time-limited token. Access renews automatically for up to 30 days, after which you sign in again.

Test your connection

Ask your assistant: "List my courses." If you see your real course names, everything works. If you get an error or nothing back, see Troubleshooting at the end of this topic.

Claude (web, desktop and mobile)

Custom connectors are available on Claude for Work (Team and Enterprise) and on individual paid plans.

Claude Team or Enterprise: an Owner adds it once

  1. Go to Organization settings and then select Connectors.

  2. Select Add, point to Custom, and then select Web.

  3. Enter a name, such as Blackboard, and the server URL, https://mcp.api.blackboard.com.

  4. Leave the OAuth client ID and secret blank, and then select Add.

Each person connects their own account

  1. In Claude, go to Customize and then select Connectors.

  2. Find Blackboard, which has a Custom label, and select Connect.

  3. Sign in with your Blackboard LMS account.

  4. In a new chat, open the tools menu and make sure Blackboard is turned on.

Set approvals. In each chat's tool settings you can choose whether Claude asks before each Blackboard action. Your Owner can set defaults for the organization. We recommend letting reads run and requiring approval for anything that changes data.

Claude Code

Run this once in a terminal:

claude mcp add --transport http blackboard https://mcp.api.blackboard.com

By default, the server is added for you in the current project only. Add the optional --scope parameter to change that:

Scope

Where the server is available

--scope local

The default. Only you, in the current project.

--scope project

Everyone working in the current project. Claude Code saves it to the project's .mcp.json file, which you can share through source control.

Each person still signs in with their own Blackboard LMS account.

--scope user

Only you, in all of your projects.

For example, to make it available in all of your projects:

claude mcp add --transport http --scope user blackboard https://mcp.api.blackboard.com

Then start Claude Code and run /mcp. Select blackboard and choose Authenticate to sign in.

To remove it later, run claude mcp remove blackboard.

ChatGPT

Custom MCP connectors that can both read and change data are available on ChatGPT Business, Enterprise and Edu. A workspace admin sets them up.

Workspace admin: create the connector

  1. Turn on developer mode for your workspace. In Workspace settings, go to Permissions & Roles then Connected Data and turn on Developer mode.

  2. In Settings and then Apps, choose to create an app or connector.

  3. Enter a name, such as Blackboard, a short description, and the MCP server URL, https://mcp.api.blackboard.com. Choose OAuth for authentication.

  4. Sign in with a Blackboard LMS test account to confirm it works, and then publish it to your workspace. On Enterprise and Edu, you can use role-based access to choose who sees it.

Each person connects their own account

  1. In ChatGPT, go to Settings and then Apps and find Blackboard, labeled Custom.

  2. Select Connect and sign in with your Blackboard LMS account.

  3. In a chat, select + and choose Blackboard so ChatGPT can use it.

Gemini

Add the Blackboard MCP Server to the Gemini app as a custom connector using only the server URL.

  1. In Gemini, go to Settings & help and then Connected apps. Select Add a custom app.

  2. Enter a name, such as Blackboard, and the server URL, https://mcp.api.blackboard.com.

  3. Sign in with your Blackboard LMS account when prompted.

Microsoft 365 Copilot

Microsoft 365 Copilot uses MCP servers through agents built in Copilot Studio. A maker or administrator builds the agent once and shares it.

  1. In Copilot Studio, create an agent or open an existing one.

  2. Go to Tools then Add a tool and then Model Context Protocol.

  3. Enter a name, such as Blackboard, and the server URL, https://mcp.api.blackboard.com. Choose OAuth 2.0 for authentication with dynamic discovery, so you don't need to enter a client ID or secret.

  4. Test the agent, and then publish it to Microsoft 365 Copilot and Teams.

  5. Each person signs in with their Blackboard LMS account the first time the agent uses a Blackboard tool.

GitHub Copilot in Visual Studio Code

  1. Open the Command Palette and run MCP: Add Server.

  2. Choose HTTP, enter https://mcp.api.blackboard.com, and name it blackboard.

  3. Choose whether to save it for this workspace or for all workspaces.

  4. Open Copilot Chat in Agent mode and select the tools icon to check that the Blackboard tools are listed. Sign in when prompted.

Or add it to .vscode/mcp.json yourself:

{
  "servers": {
    "blackboard": {
      "type": "http",
      "url": "https://mcp.api.blackboard.com"
    }
  }
}

GitHub Copilot Business and Enterprise administrators can control MCP server use with the MCP policy in their organization settings.

Codex

Add the server to your Codex configuration file, ~/.codex/config.toml:

[mcp_servers.blackboard]
url = "https://mcp.api.blackboard.com"

Then sign in and check the connection:

codex mcp login blackboard
codex mcp list

Superhuman Go by Grammarly

  1. In Superhuman Go, add a custom MCP connector with the server URL, https://mcp.api.blackboard.com.

  2. Sign in with your Blackboard LMS account when prompted.

Perplexity

  1. In Perplexity, go to Settings and then Connectors and choose to add a custom connector.

  2. Enter the server URL, https://mcp.api.blackboard.com, choose OAuth, and sign in with your Blackboard LMS account.

Perplexity doesn't accept URL parameters or custom headers. Use the address exactly as shown.

Other AI tools

Any AI tool can connect if it supports:

  • Remote MCP servers over Streamable HTTP.

  • OAuth 2.0 sign-in using the authorization code flow, with the person signing in through a browser.

  • OAuth dynamic client registration, so the tool can register itself without a client ID or secret.

Some agent platforms only support a fixed client ID and secret, or don't support browser sign-in. These can't connect.

Troubleshooting

Symptom

What to do

Your institution isn't in the list when you sign in

The server isn't turned on, or none of your host names is included in the public host lists. See Turn on the Blackboard MCP Server.

The tool shows the server as disconnected or in error

Check the URL is exactly as shown, with no extra characters. Remove the server and add it again.

Requests stop working after a while

Your sign-in has expired. Select Connect or Authenticate again, or ask your assistant to try again so it prompts you.

An action you expected isn't available

Your Blackboard LMS role may not allow it, or your AI tool's administrator may have blocked that tool.

The assistant changed something without asking

Turn on approval for write actions in your AI tool, or ask your AI tool administrator to set it for everyone.