Send to Dingding

If you're building an agent workflow for a Chinese workplace, there's a good chance the notification channel that actually matters isn't Discord or Slack - it's DingTalk (钉钉), Alibaba's enterprise messaging and office platform, roughly the regional equivalent. This node posts a fixed message straight to a DingTalk group the moment your workflow runs - no LLM decision involved, unlike the tool-calling version of this same idea (Dingding_tool).

How it works

DingTalk group chats support "custom robots" - bots you add via a webhook URL that let any external system post messages into the chat. This node is a client for that webhook: give it a URL and content, and it posts. Because it's an app-link node rather than a tool node, it fires unconditionally as a step in your graph, which makes it the right choice when you already know exactly what should be sent and when (a completion notification, a daily summary), rather than something the agent should decide about mid-conversation.

The inputs and outputs that matter

• content (default "hello world") - the message body. Wire in your LLM's output, a status string, whatever you want posted.
• msgtype (text / markdown, default markdown) - markdown gets you formatting (bold, lists, links); plain text is simpler if you just need a one-liner.
• key_word - this isn't optional busywork, it's DingTalk's own security requirement leaking through. Custom robot webhooks in DingTalk can be configured to only accept messages containing a specific keyword, as one of their built-in anti-abuse options - if your webhook has that security setting enabled, key_word needs to match, or the message gets silently rejected by DingTalk itself, not by this node.
• url (optional) - the webhook URL for your DingTalk group's custom robot.
• No outputs - this is a pure side-effect node, the end of a branch.

How to install it

• ComfyUI Manager - search "comfyui_LLM_party", install, restart.
• Manual - cd ComfyUI/custom_nodes && git clone https://github.com/heshengtao/comfyui_LLM_party, then restart.

Then pip install -r requirements.txt from inside the comfyui_LLM_party folder, in your ComfyUI Python environment (python_embeded\python.exe -m pip install -r requirements.txt on portable Windows). The real setup step happens outside ComfyUI: add a custom robot to your DingTalk group chat and grab its webhook URL - that's a DingTalk admin action, not a pip install.

Common issues & troubleshooting

Message never arrives, no error shown. Check key_word first if your DingTalk robot has keyword security enabled - a message that doesn't contain the required keyword gets rejected by DingTalk's servers, not flagged by ComfyUI. Also double-check the webhook url hasn't been rotated; DingTalk lets group admins regenerate them.

Don't leave your webhook URL sitting in a shared workflow. A DingTalk webhook URL is effectively a bearer credential for posting into your group - anyone who has it can post as your bot. If you export or share a workflow JSON that has url filled in, scrub it first, the same way you would an API key.

Wrong format shows up in the chat. If msgtype is set to text but content has markdown syntax in it, you'll see literal asterisks and pound signs in the chat rather than formatted text - match msgtype to what you're actually sending.

The pack won't load at all. Separate from this node, comfyui_LLM_party has a real, current install issue where nodes fail to import after a Manager install - usually a CUDA/PyTorch mismatch with no clear compatibility list. Check that first if Dingding isn't in your node browser.