MCP Server

The Pubble MCP (Model Context Protocol) server lets you manage your live events directly from AI assistants such as Claude. You can create and configure events, moderate audience submissions, run polls, manage your team, and analyse event performance — all from a conversational interface.

MCP (Model Context Protocol) is an open standard that lets AI assistants securely discover and call external tools and services. Your AI client acts as the MCP host, connecting to the Pubble MCP server to perform actions on your behalf.

Prerequisites

Supported Clients

The Pubble MCP server works with any client that supports Streamable HTTP transport. Verified clients include:

Connecting to the MCP Server

The Pubble MCP server is a remote HTTP server. To connect, you will need to add it to your MCP client's configuration using the following endpoint:

Replace [myteam] with your Pubble community subdomain. The subdomain scopes all tool calls to your community.

Authentication

The MCP server uses OAuth 2.0 Bearer token authentication. When your MCP client first connects, it will be redirected to the Pubble authorisation flow to grant access. Once authorised, your client will receive a Bearer token that is included automatically with each request.

If your MCP client prompts you to authenticate, sign in using your Pubble account credentials. You must be a subscriber of the community you are connecting to.

Configuring Claude Desktop

To add the Pubble MCP server to Claude Desktop, open your Claude Desktop configuration file and add the following entry under mcpServers:

{
  "mcpServers": {
    "pubble": {
      "type": "streamable-http",
      "url": "https://[myteam].pubble.io/api/v2/mcp"
    }
  }
}

Replace [myteam] with your community subdomain, then restart Claude Desktop. You will be prompted to authorise Pubble on first use.

Configuring Claude Code

To add the Pubble MCP server to Claude Code, add the following to your .mcp.json file (in your project root or home directory):

{
  "mcpServers": {
    "pubble": {
      "type": "streamable-http",
      "url": "https://[myteam].pubble.io/api/v2/mcp"
    }
  }
}

Replace [myteam] with your community subdomain. You will be prompted to authorise Pubble on first use.

What You Can Do

Once connected, you can ask your AI assistant to perform the following actions on your Pubble community:

Events

Content

Team

Analytics

Scheduling

Example Workflows

Setting up an event for a conference

Ask your assistant: "Create a Live Q&A event called 'Annual Conference 2026' starting 15 April at 9am London time, ending at 5pm"

Then: "Enable moderation for questions"

Creating an Event

Adding moderators

Ask your assistant: "Add steve@pubble.io and sarah@pubble.io as moderators"

Adding moderators

Moderating a live session

Ask your assistant: "Show me the unapproved questions for the Annual Conference event"

Then: "Approve all questions"

Approving content

Pinning a post

Ask your assistant: "Pin question [post ID] to the top"

Pinning a post

Adding a poll

Ask your assistant: "Add a new poll about how AI can be beneficial in the workplace"

Adding a poll

Post-event reporting

Ask your assistant: "Give me the stats and a summary for last week's conference Q&A"

Events reports

Permissions

The MCP server enforces the same permissions as the Pubble platform. You can only manage events and content within communities you are a subscriber of. Your OAuth token is scoped to a single community — the subdomain specified in the server URL — and grants both read and write access within that community. All tool calls are recorded in the Pubble audit log, attributed to your account.

Some destructive actions (such as deleting an event) require an explicit confirmation before they are executed. Your AI assistant will prompt you to confirm before proceeding.

Rate Limits

The MCP server applies rate limits to protect platform stability. If you exceed the allowed request rate, you will receive a 429 Too Many Requests response. Wait a moment before retrying. If you consistently hit limits during bulk operations, break your requests into smaller batches.

Troubleshooting

If you continue to experience issues, contact us using the button above or visit your Pubble dashboard for further support.