MCP server

Deckbell runs an MCP (Model Context Protocol) server, so AI assistants such as Claude and Cursor can work with your Deckbell workspace: find the issues assigned to you, create and update issues, comment, and post in channels.

• Server URL: https://api.deckbell.com/mcp (Streamable HTTP)
• Sign-in: an API token, sent as a Bearer token

The quickest way to set it up is Integrations → MCP server in the app: it makes a token for your assistant and fills it into ready-to-paste setup for each client.

Create a token for your assistant

In Integrations → MCP server choose Create a token for MCP, or make one in Settings → API tokens. The assistant acts as you, so it can only do what the token's scopes and your role allow; tick only Read scopes if you just want it to look things up. Give it its own token so you can revoke it separately.

Claude Code

Run this in a terminal, with your token:

claude mcp add --transport http deckbell https://api.deckbell.com/mcp \
  --header "Authorization: Bearer dkb_your_token"

Claude Desktop

Open Settings → Developer → Edit config, add Deckbell to claude_desktop_config.json, and restart Claude. This uses mcp-remote, so it needs Node.js installed:

{
  "mcpServers": {
    "deckbell": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://api.deckbell.com/mcp", "--header", "Authorization:${DECKBELL_AUTH}"],
      "env": { "DECKBELL_AUTH": "Bearer dkb_your_token" }
    }
  }
}

Cursor

Add this to ~/.cursor/mcp.json (or .cursor/mcp.json in a project):

{
  "mcpServers": {
    "deckbell": {
      "url": "https://api.deckbell.com/mcp",
      "headers": { "Authorization": "Bearer dkb_your_token" }
    }
  }
}

Other MCP clients that support remote (Streamable HTTP) servers with custom headers work the same way: the URL above and an Authorization: Bearer header.

Tools

Tool	What it does	Needs scope
whoami	Which person, workspace and scopes the connection uses	any
list_projects	The workspace's projects and their keys	projects:read
list_statuses_and_labels	A project's statuses and labels	projects:read or issues:read
search_issues	Search and filter issues	issues:read
list_my_issues	Your open issues (or all, with done ones)	issues:read
get_issue	An issue's details and comments	issues:read
create_issue	Create an issue	issues:write
update_issue	Change status, priority, assignee, labels, title, description, estimate; archive	issues:write
add_comment	Comment on an issue, as you	comments:write
list_channels, read_channel	Channels you're in and their recent messages	chat:read
post_message	Post in a channel, as you	chat:write
list_members	People in the workspace	members:read

The assistant only sees the tools its token can use.

Things to ask

• "What's assigned to me in Deckbell?"
• "Create a bug in WEB: the checkout button does nothing on Safari. High priority."
• "Move WEB-42 to In Progress and add a comment that I'm on it."
• "Summarize what was said in #general today."

Comments and messages the assistant writes appear under your name, and mentioning an AI agent ("@Coder take this") in a comment starts it, as in the app. Calls count towards the token's rate limit. Revoke the token in Settings → API tokens to disconnect the assistant.