Dandex

Set Up Dandex

Add one server URL. Your assistant then opens dandex.ai to sign in, where you paste your API token once. The token is in your welcome email.

https://dandex.ai/mcp

No token yet? Get one by email. Setting up with an agent? Give it this prompt:

Chat assistants

  • Claude

    claude.ai, Desktop and mobile, on every plan (Free allows one custom connector). On Team and Enterprise an owner adds it first.

    Add to Claude

    Add for your organization (owners)

    Manual setup
    1. Open the link. Claude shows "Add custom connector" with Dandex and its URL filled in. Press Add.
    2. Press Connect. Claude opens dandex.ai; paste your API token and press Authorize.
    3. Dandex now works in Claude on the web, Desktop and mobile, on the same account.

    Claude documentation

  • ChatGPT

    No pre-fill link exists. If Plugins has no "Add custom MCP server", your plan or workspace does not allow it.

    Open ChatGPT plugins

    Copies the server URL

    Manual setup
    1. On the Plugins page, press + and choose Add custom MCP server.
    2. Name it Dandex, paste https://dandex.ai/mcp as the URL and choose OAuth.
    3. Tick "I understand and want to continue", press Create as a plugin, then paste your API token on dandex.ai and press Authorize.
    4. In a chat, type @ and pick Dandex.

    ChatGPT documentation

  • Mistral Le Chat

    Only workspace administrators can add connectors. On Free and Pro you are the administrator.

    Manual setup
    1. Open Connectors, press + Add Connector and choose the Custom MCP Connector tab.
    2. Name it dandex (no spaces) and paste https://dandex.ai/mcp.
    3. Press Connect. Le Chat opens dandex.ai; paste your API token and press Authorize.

    Mistral Le Chat documentation

  • Perplexity

    Pro and above. On Enterprise an administrator controls connectors.

    Manual setup
    1. Open Settings, then Connectors, and press + Custom connector.
    2. Choose Remote, paste https://dandex.ai/mcp, and pick OAuth and Streamable HTTP.
    3. Perplexity opens dandex.ai; paste your API token and press Authorize.

Code editors

  • Cursor

    Opens Cursor with Dandex ready to install.

    Add to Cursor

    Open in Cursor

    Manual setup
    1. Press Add to Cursor and confirm the install in Cursor.
    2. In Settings, Tools & MCP, press the sign-in prompt next to dandex.
    3. Cursor opens dandex.ai; paste your API token and press Authorize.

    Or add it by hand

    ~/.cursor/mcp.json

    {
  "mcpServers": {
    "dandex": {
      "url": "https://dandex.ai/mcp"
    }
  }
}

    Cursor documentation

  • VS Code

    GitHub Copilot agent mode.

    Add to VS Code

    Open in VS CodeVS Code Insiders

    Manual setup
    1. Press Add to VS Code and choose Install in the MCP server view.
    2. VS Code asks to sign in to dandex; allow it.
    3. Paste your API token on dandex.ai and press Authorize.

    Or from a terminal

    code --add-mcp '{"name":"dandex","type":"http","url":"https://dandex.ai/mcp"}'

    VS Code documentation

  • Windsurf

    Now Devin Desktop (Cascade). No install link exists.

    Manual setup
    1. Open the Cascade MCP config file shown below.
    2. Add the dandex entry and save.
    3. Refresh MCP servers in Cascade. It opens dandex.ai; paste your API token and press Authorize.

    Config

    ~/.config/devin/mcp_config.json (Windows: %APPDATA%\devin\mcp_config.json)

    {
  "mcpServers": {
    "dandex": {
      "serverUrl": "https://dandex.ai/mcp"
    }
  }
}

    Windsurf documentation

  • Zed

    Settings, AI, MCP Servers, Add Remote Server. Or paste the config.

    Manual setup
    1. Open Settings, then AI, then MCP Servers, and press Add Remote Server.
    2. Name it dandex and paste https://dandex.ai/mcp. Leave headers empty.
    3. Zed opens dandex.ai; paste your API token and press Authorize.

    Or in settings.json

    settings.json

    {
  "context_servers": {
    "dandex": {
      "url": "https://dandex.ai/mcp"
    }
  }
}

    Zed documentation

  • Cline

    Set type to streamableHttp, or Cline falls back to SSE.

    Sign-in with dandex.ai is not confirmed for this client. If it does not open, use the header config under Clients without sign-in.

    Manual setup
    1. In Cline, open MCP Servers and choose Configure MCP Servers.
    2. Add the dandex entry and save.

    Config

    cline_mcp_settings.json

    {
  "mcpServers": {
    "dandex": {
      "type": "streamableHttp",
      "url": "https://dandex.ai/mcp"
    }
  }
}

    Cline documentation

Command line

  • Claude Code

    Then run claude mcp login dandex to sign in.

    claude mcp add --transport http --scope user dandex https://dandex.ai/mcp
    Manual setup
    1. Run the command in a terminal.
    2. Run claude mcp login dandex. It opens dandex.ai; paste your API token and press Authorize.
    3. Or, inside Claude Code, run /mcp, select dandex and choose Authenticate.

    Claude Code documentation

  • Gemini CLI

    Then run /mcp auth dandex inside Gemini CLI.

    gemini mcp add --scope user --transport http dandex https://dandex.ai/mcp
    Manual setup
    1. Run the command in a terminal.
    2. Start Gemini CLI and run /mcp auth dandex.
    3. Paste your API token on dandex.ai and press Authorize.

    Gemini CLI documentation

  • Codex CLI

    Then run codex mcp login dandex.

    codex mcp add dandex --url https://dandex.ai/mcp
    Manual setup
    1. Run the command in a terminal.
    2. Run codex mcp login dandex.
    3. Paste your API token on dandex.ai and press Authorize.

    Or in config.toml

    ~/.codex/config.toml

    [mcp_servers.dandex]
url = "https://dandex.ai/mcp"

    Codex CLI documentation

Desktop apps

  • Goose

    Opens Goose with Dandex ready to add. Sign-in for remote extensions is in beta.

    Sign-in with dandex.ai is not confirmed for this client. If it does not open, use the header config under Clients without sign-in.

    Add to Goose
    Manual setup
    1. Press Add to Goose and confirm in Goose.
    2. If Goose opens dandex.ai, paste your API token and press Authorize.
    3. If it does not, add the header from "Clients without sign-in" to the extension.

  • LM Studio

    Opens LM Studio with Dandex ready to add.

    Sign-in with dandex.ai is not confirmed for this client. If it does not open, use the header config under Clients without sign-in.

    Add to LM Studio
    Manual setup
    1. Press Add to LM Studio and confirm in LM Studio.
    2. If LM Studio does not open dandex.ai to sign in, add the Authorization header in its mcp.json.

    Header fallback

    mcp.json

    {
  "mcpServers": {
    "dandex": {
      "url": "https://dandex.ai/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_DANDEX_TOKEN"
      }
    }
  }
}

    LM Studio documentation

Any other MCP client

Name
dandex
Transport
Streamable HTTP
URL
https://dandex.ai/mcp

mcp.json

{
  "mcpServers": {
    "dandex": {
      "url": "https://dandex.ai/mcp"
    }
  }
}

Clients without sign-in

If your client cannot open dandex.ai to sign in, send your API token in the Authorization header. Never put it in the URL: URLs end up in logs and browser history.

mcp.json

{
  "mcpServers": {
    "dandex": {
      "url": "https://dandex.ai/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_DANDEX_TOKEN"
      }
    }
  }
}

Replace YOUR_DANDEX_TOKEN with the token from your welcome email. Lost it? Enter the same email below to issue a new one.

Ask your first question

Once connected, ask your assistant:

The Dandex skill teaches your assistant longer workflows: credit checks, due diligence reports and sales prospecting.

Scheduled tasks in Claude Desktop

Claude Desktop runs scheduled tasks against local MCP servers only, so the connector alone is not enough. Add Dandex through mcp-remote as well. The first time it starts, it opens dandex.ai to sign in, the same as the connector.

  1. In Claude Desktop, open Settings, then Developer, and press Edit Config.
  2. Add the dandex entry below and save.
  3. Check Node.js 18 or later is installed (node --version), then restart Claude Desktop.

claude_desktop_config.json

{
  "mcpServers": {
    "dandex": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://dandex.ai/mcp"
      ]
    }
  }
}

Then ask Claude to schedule a task, for example:

REST API and command line

Every tool is also a JSON endpoint. POST its input to /api/v1/tools/<name> with your API token in the Authorization header; the token and quota are the ones your assistant uses. Set DANDEX_TOKEN to your token first.

curl -s https://dandex.ai/api/v1/tools/company-overview \
  -H "Authorization: Bearer $DANDEX_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"cvrNumber":"35527710"}'

The OpenAPI document describes every tool, its input and its errors.

The dandex command, a Node.js CLI, runs the same calls from a terminal. It reads the token from DANDEX_TOKEN or from dandex login.

npm install -g dandex
dandex login
dandex lookup 35527710
dandex search "Novo Nordisk" --limit 5
dandex tools

Get your API token

Enter your email and the token arrives in your inbox. Already signed up? The same form emails you a link to issue a new one.

Free tier: 100 requests/day. No credit card required.

Or try the chat demo first