IrodoriTTS-v2 Emoji Selector

IrodoriTTS-v2 Emoji Selector has no inputs and no outputs. It does nothing to your audio. Which makes it sound like a waste of a node slot - until you realize what it's for: IrodoriTTS-v2 understands emoji embedded in the text as emotional and speech-effect tags, and this node is a cheat sheet that lets you copy them.

The web extension that powers this node renders a panel of buttons, one per supported emoji, each labeled in Japanese with what it does. The list is worth reading once because it's the model's built-in expressiveness vocabulary - a partial sample:

• 😭 - sobbing, crying, grief
• 📢 - echo / reverb (i.e., "through a megaphone")
• 🥵 - panting, moaning, groaning
• ⏸️ - pause, silence
• 🐢 - slowly
• ⏩ - fast, rushing
• 🫶 - gently
• 👂 - whispering, close to the ear
• 😮💨 - sighing, breathing

Click an emoji in the panel and it's copied to your clipboard; paste it into the text field of an IrodoriTTS-v2 Sampler, or into the text of a line in your dialogue_segments_json, and the model tries to deliver that line with the matching delivery. This is genuinely one of the nicer parts of IrodoriTTS-v2 - fine-grained performance control that rides inside the plain text, no extra nodes or conditioning required.

The workflow it pairs with best: you're writing a two-character scene in the Dialogue TTS node, and one line needs to be a whispered aside or a sudden shout. Drop the emoji into that utterance's text and the performance lands without you re-cloning a voice or adding post-processing.

Honest caveats. The emoji are guidance, not guarantees - a single 📢 isn't going to turn a line into a full studio reverb effect, and the model is Japanese-first, so the emotional tags sit naturally in Japanese text. And because the labels live in Japanese, a quick look at the list (or this page) saves you guessing what 🥴 means.

Everything else about it is the pack: it installs with ComfyUI-DialogueTTS via ComfyUI Manager (search "ComfyUI-DialogueTTS") or

cd ComfyUI/custom_nodes
git clone https://github.com/kantan-kanto/ComfyUI-DialogueTTS
pip install -r ComfyUI-DialogueTTS/requirements.txt

then restart, drop the checkpoint in models/checkpoints/ (it won't download itself), and let the codec and tokenizer fetch on first load. If you've never seen the emoji panel after installing, hard-refresh the ComfyUI browser tab - the extension lives in web/, and a stale frontend is the usual reason a widget doesn't show up.