Playbook Text (External)

Playbook Text is a multiline string passthrough with a second output for trigger words. You type a prompt in, you get a prompt out - plus a separate channel for the LoRA trigger words that should ride along with it. It's one of the least magical nodes in the pack, and that's fine, because its job is plumbing.

This is an "External" node in Playbook Nodes, the ComfyUI side of Playbook3D's cloud render service. On their platform, the web UI drives these nodes - a text field in their interface becomes a prompt in your workflow, and the id/label fields are the binding keys. Locally, it's a convenience: one node that carries both the prompt and its trigger words, so a workflow stays tidy when you're passing styled text around.

The inputs and outputs

• id and label - strings, the platform-binding metadata. Inert locally.
• default_value - the prompt, multiline.
• trigger_words - a second multiline string, which is where the pack's intent shows: LoRAs trained on characters or styles typically come with a trigger phrase that unlocks them, and keeping it in a dedicated field means you can stitch it into your prompt reliably instead of remembering to type it every time.

Outputs are text (equal to default_value) and trigger_words (equal to that field). The node literally returns what you gave it - parse_text is a two-line identity function. It makes no network calls, needs no API key, and will happily output empty strings if you feed it nothing.

Install

The standard pack install:

cd ComfyUI/custom_nodes
git clone https://github.com/playbook3d/playbook3d-comfyui-nodes

Or search Playbook Nodes in ComfyUI Manager, then restart ComfyUI. No models, no dependencies beyond ComfyUI's built-ins.

Troubleshooting

The honest gotcha is that this node validates nothing. It won't strip formatting, it won't reject empty strings, and it certainly won't check whether your trigger words match a LoRA you own. If your prompt output looks wrong, you're probably looking at a case where default_value and trigger_words got mixed up in the wiring - wire text into your CLIP Text Encode and trigger_words into whatever composes the final prompt, and keep them straight. One more thing: if you're used to ComfyUI's built-in text nodes, this one's only real advantage is the second output. If you don't need trigger words split out, the core nodes do the same job with less pack-specific lock-in.