Connect your AI assistant (MCP)

Let Claude and other MCP clients create, edit and organise your boards in plain conversation

The Model Context Protocol (MCP) is how AI assistants such as Claude connect to external tools. There are two ways in: claude.ai can talk to the endpoint we host (nothing to install), or a desktop client such as Claude Desktop, Claude Code or Cursor can run our small server locally. Both expose the same tools and both work through the REST API, so an assistant can only see and change what your API key allows, and everything it does fires your board automations exactly like an edit made in the app.

1. Create an API key

Go to Profile > API keys and create a key. Copy the ck_... value when it is shown: it is only displayed once. API access is included from the Team plan.

2. On claude.ai (nothing to install)

We host an MCP endpoint with full OAuth support, so claude.ai on the web or mobile can connect straight to it. In claude.ai go to Settings > Connectors > Add custom connector and paste this URL:

/api/mcp

When you first use the connector, your browser opens this site so you can sign in and approve the connection. Approving creates an API key for it automatically; you can revoke that key at any time under Profile > API keys, which disconnects the assistant. You do not need to create a key by hand for this route, so step 1 is optional here. Other MCP clients that support remote servers can use the same URL, either via the same OAuth flow or with Authorization: Bearer ck_your_key set directly.

3. Or download the server for desktop clients

You need Node.js 18 or newer. In a terminal:

mkdir conneqtcrm-mcp && cd conneqtcrm-mcp
curl -O /mcp/conneqtcrm-mcp.mjs
npm install @modelcontextprotocol/sdk zod

4. Add it to your desktop assistant

Replace ck_your_key with your API key and the path with wherever you downloaded the file.

Claude Code

claude mcp add conneqtcrm \
  -e CONNEQTCRM_API_URL= \
  -e CONNEQTCRM_API_KEY=ck_your_key \
  -- node /path/to/conneqtcrm-mcp/conneqtcrm-mcp.mjs

Claude Desktop

Settings > Developer > Edit Config, then add this to claude_desktop_config.json and restart Claude Desktop:

{
  "mcpServers": {
    "conneqtcrm": {
      "command": "node",
      "args": ["/path/to/conneqtcrm-mcp/conneqtcrm-mcp.mjs"],
      "env": {
        "CONNEQTCRM_API_URL": "",
        "CONNEQTCRM_API_KEY": "ck_your_key"
      }
    }
  }
}

Cursor and other MCP clients accept the same command, arguments and environment variables in their MCP settings.

5. Try it

Ask your assistant something like "List my boards", "Add a lead called Acme to my Lead Tracker with status Contacted" or "Post a comment on the Acme lead saying the proposal was sent".

What the assistant can do

list_boards / get_board

Find boards, columns, groups and members

create_board / create_group

Set up new boards and sections

list_items / get_item

Read items with all column values

create_item

Add items, with column values, in one step

update_item / delete_item

Rename items or move them to the recycle bin

set_item_value

Change a status, date, text or person value

upsert_items

Bulk import that updates existing items by a match column

list_comments / add_comment

Read and post on an item's thread

Security notes

The assistant acts as you: it can only reach boards you can, within the key's scopes and rate limit.
Treat the API key like a password. You can revoke it at any time from Profile > API keys.
Deletions are soft: items removed by the assistant sit in the board's recycle bin until restored or purged.

Prefer raw HTTP?Read the REST API docsor seeall integrations and webhooks