DevvyBoard

Team workspace

Developer docs

MCP Integration

Connect AI coding tools to your DevvyBoard projects. With the Model Context Protocol integration, Cursor, Claude Code, and Codex CLI can list projects, read boards, and create or move cards without leaving your editor.

Quick start

New to MCP? Follow these steps in order. The whole setup usually takes about five minutes.

1

Sign in to DevvyBoard — Free or Pro (no paid plan required for MCP).

2

Click your profile picture in the top-right corner and choose Connectivity. You can also open Connectivity from the mobile menu.

3

Click Create API key, name it (for example, Cursor on laptop), pick the projects to share, and choose read or read-and-write access. Copy the key when shown — it starts with dbk_live_ and appears only once.

4

Add the DevvyBoard MCP server to your AI tool using the config in the Cursor, Claude Code, or Codex section below. Paste your API key where it says dbk_live_YOUR_KEY_HERE.

5

Restart or reload MCP in your tool, then ask the AI to “list my DevvyBoard projects” to confirm everything works.

The MCP package runs with one command — no clone or build step: npx -y devvyboard-mcp (published as devvyboard-mcp@0.1.1 on npm).

Overview

DevvyBoard exposes a Model Context Protocol (MCP) server that AI assistants can call as tools. Once configured, your AI assistant can:

  • List all projects your API key is authorised for
  • Read the full Kanban board — columns and cards — for any project
  • Fetch individual card details including description, priority, and deadline
  • Create cards in any column (write scope required)
  • Update card title, description, priority, or deadline (write scope required)
  • Move a card to another column (write scope required)
  • Delete a card (write scope required)
The MCP package is published as devvyboard-mcp@0.1.1 on npm. It runs as a local stdio server with npx -y devvyboard-mcp — no additional service or infrastructure is required on your machine.

Prerequisites

1

DevvyBoard account (Free or Pro). Free includes MCP with daily API limits; Pro raises limits for heavy automation.

2

Node.js 20 or later installed on the machine running your AI tool. The MCP server is distributed as an npm package and executed with npx.

3

An API key generated from Connectivity (profile menu → Connectivity). See the next section.

Get an API key

API keys are scoped to specific projects and carry either read or read-and-write permissions. You can create and revoke keys at any time.

1

Sign in to DevvyBoard on web. Click your profile picture (top-right) and choose Connectivity. Or go directly to /connectivity.

2

Click Create API key.

3

Give the key a name (for example, Cursor MCP), select the projects to grant access to, and choose whether the key should have read or read and write scope.

4

Click Create. Copy the key shown — it begins with dbk_live_ and is displayed only once.

Store the key in your OS keychain or password manager. If you lose it, revoke the key from the Connectivity page and create a new one.

Cursor

Add the DevvyBoard MCP server to your Cursor configuration. Cursor reads MCP servers from ~/.cursor/mcp.json (global) or a .cursor/mcp.json file at the root of any project (project-scoped).

{
  "mcpServers": {
    "devvyboard": {
      "command": "npx",
      "args": ["-y", "devvyboard-mcp"],
      "env": {
        "DEVVYBOARD_API_KEY": "dbk_live_YOUR_KEY_HERE"
      }
    }
  }
}
1

Open or create the file at ~/.cursor/mcp.json (on Windows: %USERPROFILE%\.cursor\mcp.json).

2

Paste the configuration above, replacing dbk_live_YOUR_KEY_HERE with your actual API key from Connectivity.

3

In Cursor, open the Command Palette (Ctrl+Shift+P on Windows/Linux, Cmd+Shift+P on Mac) and run MCP: Restart all servers.

4

The devvyboard server and its tools should appear in the MCP panel. Ask the AI agent to “list my DevvyBoard projects” to confirm connectivity.

Claude Code

Claude Code uses a claude_desktop_config.json file for MCP configuration. Add the DevvyBoard server under the mcpServers key.

Config file location by platform:

  • macOS / Linux: ~/Library/Application Support/Claude/claude_desktop_config.json
  • Windows: %APPDATA%\Claude\claude_desktop_config.json
{
  "mcpServers": {
    "devvyboard": {
      "command": "npx",
      "args": ["-y", "devvyboard-mcp"],
      "env": {
        "DEVVYBOARD_API_KEY": "dbk_live_YOUR_KEY_HERE"
      }
    }
  }
}
1

Open or create the config file at the path above for your platform.

2

Add the devvyboard entry under mcpServers, replacing the placeholder key.

3

Restart Claude Code. The DevvyBoard tools will appear in the tools panel.

Codex CLI

Codex CLI supports MCP servers via a codex.json configuration file in your project root or home directory.

{
  "mcpServers": {
    "devvyboard": {
      "command": "npx",
      "args": ["-y", "devvyboard-mcp"],
      "env": {
        "DEVVYBOARD_API_KEY": "dbk_live_YOUR_KEY_HERE"
      }
    }
  }
}

Alternatively, export the key as an environment variable before running Codex:

export DEVVYBOARD_API_KEY=dbk_live_YOUR_KEY_HERE
codex
Refer to the Codex CLI documentation for the exact config file path on your version.

Available tools

The MCP server exposes the following tools. Each tool call targets the DevvyBoard REST API on your behalf using the API key supplied in your configuration.

ToolDescriptionMin. scope
list_projectsList all projects accessible with this API key.read
get_boardReturn the columns and cards for a project.read
get_cardFetch full details for a single card by ID.read
create_cardAdd a new card to a specified column.write
update_cardChange a card's title, description, priority, or deadline.write
move_cardMove a card to a different column at a given position.write
delete_cardPermanently delete a card from the board.write

Scopes

Every API key carries one of two scope levels. You choose the scope when creating the key from Connectivity.

Read

The key can list projects, read boards, and fetch individual cards. It cannot make any changes.

Suitable for AI assistants that summarise work or answer questions about tasks.

Read and Write

The key can read everything plus create, update, move, and delete cards.

Suitable for AI agents that manage your board on your behalf.

Prefer read-only keys when an AI tool only needs to look up tasks. Limit write access to agents you fully trust to modify your board.

API limits

Each HTTP call under /api/integrations/v1/* counts toward your quota (MCP tools map one-to-one). Exceeding a limit returns HTTP 429 with rate-limit headers — not a silent throttle.

LimitFreePro (incl. legacy Team)
Active API keys210
Daily API requests30010,000
Daily writes502,000
Burst (all methods)20 / minute120 / minute
Scopesread + writeread + write
Pro is for higher headroom, not exclusive MCP access. Upgrade from billing in the mobile app when you need more daily requests.

FAQ

My AI tool returns a 401 error — what should I check?

Verify that the API key is copied correctly and has not been revoked from Connectivity. Ensure the key prefix is dbk_live_ and the full value is passed as the DEVVYBOARD_API_KEY environment variable.

Can I use one key for multiple projects?

Yes. When creating a key, select all the projects you want it to access. The list_projects tool will return only those projects.

Can I scope a key to read-only to prevent accidental changes?

Yes. Select read-only scope when generating the key. Any tool that requires write access will return a permission error, leaving your board unchanged.

How do I rotate a compromised key?

Open Connectivity from your profile menu, revoke the old key, then create a new one. Update your MCP configuration with the new key and restart your AI tool.

Are there rate limits on API key usage?

Yes. Free accounts get 300 API requests per day (50 writes) and a per-minute burst cap. Pro raises those limits. When exceeded, the API returns HTTP 429 and your MCP client shows an error. See API limits above or upgrade to Pro for heavier automation.

Do I need Pro to use MCP?

No. Free accounts can create API keys from Connectivity and use all MCP tools with read and write scopes, subject to daily limits. Pro is optional for higher quotas.

Does the MCP server persist any data locally?

No. The MCP package is a stateless stdio server. All data lives in your DevvyBoard workspace and is fetched or written in real time through the API.

Ready to connect your AI tool?

Sign in to DevvyBoard (Free or Pro) and generate your first API key from Connectivity.

Developer docs | DevvyBoard