Text Input v2

This is the simplest node in the whole Searge SDXL pack, and the one you'll actually type into. "Text Input v2" is just a multiline text box: you write a prompt, and it emits that text in the format the rest of Searge's prompting stages expect. That last part is the only reason it exists as a separate node instead of you using a generic string primitive.

Searge SDXL's prompting runs on its own text type, SRG_PROMPT_TEXT, so the prompt data can travel cleanly into the prompt adapter and onto the Magic Box data stream. SeargeTextInputV2 is the node that produces that type from plain typed text. Think of it as the on-ramp: your words go in here, and everything downstream in the Searge prompting chain speaks the same dialect.

How it works

You get a text field. You type. It outputs your text as an SRG_PROMPT_TEXT. There's no encoding, no CLIP, no conditioning happening here - this is pure text capture. The actual prompt encoding happens later in the pipeline; this node's job is just to hold what you wrote and pass it along in the right wrapper.

In a normal Searge setup you'll have several of these - one for your main positive prompt, one for the negative, and more if you're using the secondary and style lanes that SDXL's dual text encoders make available. Each text box feeds a corresponding slot on the prompt adapter. Keeping the text in its own node like this is what lets Searge give you those separate, independently-editable prompt lanes instead of one cramped box.

The inputs and outputs that matter

Nothing could be leaner:

• prompt (required STRING, multiline, default empty) - the text field. This is where you write the prompt. That's the whole input surface.
• output prompt_text (SRG_PROMPT_TEXT) - your text, wrapped in Searge's prompt type. Wire it into a slot on SeargePromptAdapterV2.

How to install it

Ships with the Searge SDXL pack. ComfyUI Manager: search SeargeSDXL, install, restart. Or:

cd ComfyUI/custom_nodes
git clone https://github.com/SeargeDP/SeargeSDXL.git

then restart. Manual installers need to run python -m pip install opencv-python once in ComfyUI's Python environment or the pack won't load.

Common issues

The one thing that trips people is the output type. prompt_text is SRG_PROMPT_TEXT, a Searge-specific type - it only plugs into Searge prompting nodes, not into a stock CLIP Text Encode. If you're trying to feed a Searge prompt into a non-Searge node, that mismatch is why the wire won't connect; you'd need to route it through Searge's own encoding path. Otherwise there's almost nothing to go wrong here. As with the rest of the pack, if the node vanishes or errors after a ComfyUI update, that's a pack-loading issue - update Searge (the v4.3.x releases were compatibility fixes) and keep the workflow JSON matched to the installed node version.