Prompt Generator
Add Node → Prompt → Prompt Generator
A simple chat with an LLM whose only job is writing prompts for your generators. Describe what you’re after in plain language, and it hands you back a finished prompt you can wire straight into an Image or Video Generator.
You can also connect images to it — up to ten. That lets you ask things like “write me a prompt that matches this style” or “describe this scene,” with the model actually looking at what you gave it.
What’s on the node
| Area | What goes in it |
|---|---|
| Conversation | Everything the model has written so far. Click any earlier reply to make it the active one again. |
| Chat box | Where you type the brief, in plain language. |
| Model | Which model answers. |
| Language | What language the prompt comes back in. English by default, whatever language you wrote your brief in — set it to Auto to get the prompt in your own language instead. |
| Output | How many prompts one brief returns: 1, 2, 4, 6, 8 or 10. Anything above 1 adds that many output handles — see below. |
| I | Up to ten images for the model to look at while it writes. |
| P | The outputs, on the right edge: one numbered ring per prompt, each wired to its own generator. |
A ring is hollow until something is connected and fills with colour once it is. There’s more on all of this in How nodes connect.
One brief, several prompts
The Output control is the part worth understanding. Leave it at 1 and the node behaves like a normal chat: one brief in, one prompt out, one P handle on the right.
Raise it to 2, 4, 6, 8 or 10 and the node grows that many output handles, numbered down the right edge. You still write a single brief — you don’t describe each variation separately — and the model turns it into that many different prompts, each its own take on the same idea rather than a copy of it. Handle 1 carries the first, handle 2 the second, and so on.
Wire each handle to its own generator and one sentence from you becomes ten images exploring the same brief in ten directions, with nothing to copy-paste. Bear in mind every handle is a generation of its own, so ten outputs cost ten generations.
Which model should you use?
| Model | Sees images | What it’s good at | |
|---|---|---|---|
| Pro | ChatGPT 5.5 | yes | The default, and a strong general choice. Writes solid, natural prompts and reads whatever images you connect. |
| Good | Qwen 3.6 Plus | yes | Lighter and cheaper. Handy when you’re iterating quickly and don’t need the extra polish. |
| Creative | Qwen3.5 27B | yes | The smallest of the three. Fine for straightforward rewrites and quick variations on a brief. |
Prices are indicative and set by the providers, who revise them: see wavespeed.ai or runware.ai for current rates.
Prompts cost a fraction of a generation — around a cent a reply, or five cents when the model has to look at connected images — so the model you pick here matters far less to your bill than the one on the generator downstream.
A couple of useful things
Click any earlier reply in the conversation to make it the active one again — the connected handle switches to that prompt instantly, so you can go back to a better answer without redoing anything.
If you want finer control over how a multi-prompt reply gets split up and routed, pair this node with Prompt Split.