💬 Auto-MSG-ALL

Auto-MSG-ALL is the master switch of xlinx's auto-messaging pack: instead of picking one chat service, it can fire your finished render into LINE, Telegram, and Discord from a single node. Same core machinery as the sibling nodes - IMAGE in, the image gets POSTed over HTTP, and a slice of your prompt rides along in the caption - just with every option on the table at once.

Why would you want that? Maybe you run a couple of machines and want results in more than one place: Telegram for you, a Discord channel for the team or a client, LINE for the person who insists on LINE. The author is Taiwan-based, which is exactly why LINE gets equal billing here when most Western packs ignore it. If you're juggling multiple services, one node beats three.

The inputs you actually set

ALL exposes everything, and that's both the point and the trap. You get:

• Three enable toggles: enable_im_line_notify, enable_im_telegram, enable_im_discord
• Three token/chat-ID pairs, one per service - im_line_notify_token, im_telegram_token_botid + im_telegram_token_chatid, im_discord_token_botid + im_discord_token_chatid
• Three message headers, one per service
• The same msg_png_info_enabled / msg_pos_prompt_enabled / msg_neg_prompt_enabled toggles shared by the whole pack - which, in the code as shipped, are accepted but never actually change anything. The prompt is always appended.

That's thirteen required inputs before you even touch the pass-throughs. The trigger_any_type and trigger_append_image optional inputs work exactly like the other nodes: connect your final output and your VAE-decoded image so the node can both send and stay inline in the graph.

Outputs are the same pass-through pair plus three STRING outputs - 🌀output-text-LINE, 🌀output-text-Telegram, 🌀output-text-Discord. Each carries that service's raw API response, which is how you tell which channel worked and which one 400'd.

Where people get burned

Don't use this node if you only use one service. The per-channel nodes (Auto-MSG-Telegram-Bot, etc.) hide the other two services' fields entirely; ALL shows every empty token box and it's too easy to leave a stray setting half-filled. For one channel, use the dedicated node. For multi-channel, ALL is genuinely convenient.

Install

ComfyUI Manager: search "ComfyUI-decadetw-auto-messaging-realtime". Or:

cd ComfyUI/custom_nodes
git clone https://github.com/xlinx/ComfyUI-decadetw-auto-messaging-realtime

Restart ComfyUI. No requirements.txt, no model downloads - just requests and Pillow, both already present in a stock ComfyUI.

Common issues

• One channel silent while others work? Check that channel's STRING output for the API error, and double-check its enable toggle - a token is irrelevant if the toggle is off.
• All your tokens live as plaintext widgets in the workflow JSON. Never share a workflow without scrubbing them; bots get hijacked this way. The pack also deliberately ships without any "send commands back to ComfyUI" feature - the author killed two-way messaging over security concerns, so this is strictly one-way.