Send Message to Discord
Post text, images, or audio from your workflow
Once your agent workflow has produced something worth sharing - an LLM's answer, a status update, maybe an image from dall_e_tool - you need a way to actually put it in front of someone. Send Message to Discord is the direct-send half of comfyui_LLM_party's Discord integration: it posts text, and optionally an image and audio file, to Discord the moment your workflow runs.
How it works
This runs through the same live connection discord_bot opens, but where discord_file_monitor reads from Discord, this node writes to it. It's an unconditional step - wire it into your graph and it fires every run, unlike a tool-calling node where the LLM decides whether to act. That makes it the right choice for "always tell me the result" workflows, and the wrong choice if you actually want the agent judging whether a message is worth sending (there's no tool-calling equivalent of this specific node in the pack the way there is for DingTalk).
The inputs and outputs that matter
text- the message body.img_path- a local file path to an image to attach, if you have one (say, output from an image-generation branch of your workflow).audio_path- same idea, for an audio file - useful if your workflow includes a text-to-speech step and you want the agent's response delivered as a voice clip rather than just text.is_enable- toggle sending off without removing the node.- No outputs - this is a pure side-effect node, the end of a branch.
Since img_path and audio_path point at local files rather than embedding the content directly, pairing this node with clear_file afterward is a sensible way to clean up the temp file once it's been sent.
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). This node needs discord_bot connected first - same Discord Developer Portal setup (application, bot token, invited to your server with send-message permission) applies here, since there's no separate credential on this node itself.
Common issues & troubleshooting
Message never arrives. Check discord_bot's connection status first - this node has nothing to send through if the bot never authenticated. After that, verify the bot actually has send-message permission in the specific channel you're targeting; being in a server doesn't guarantee permission in every channel.
Attachment doesn't show up. img_path and audio_path need to be valid local paths that exist at the moment this node runs - if an upstream generation step hasn't finished writing the file yet, or the path is wrong, the attachment silently fails to send even if the text goes through.
Files piling up on disk. Since attachments go by file path rather than embedded data, nothing here deletes the source file after sending. Wire the output of your generation step through clear_file after this node if you don't want temp images and audio accumulating.
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 discord_send isn't in your node browser.
Inputs (4)
| Name | Type | Default | Description |
|---|---|---|---|
| is_enable | BOOLEAN | true | — |
| text | STRING | — | |
| img_path | STRING | — | |
| audio_path | STRING | — |
Outputs (0)
No outputs