Spark

In the style profile, your whole style sits in ONE field, “STYLE PROMPT”: persona, tone rules, stage logic and forbidden phrasings, in whatever order you like. With “Generate”, Spark turns it into sentences for each stage, which you retrieve by voice command. The { } icon at the top right gives you the configuration as JSON, which you can copy into any language model — it explains everything to you and, if you like, builds a ready-made profile for your case. Anyone can view it. Copying and saving as a file (“Copy”, “Save (.json)”) is part of the Pro plan (Agency too). There are two modes — chat and style profile — and a temperature setting. You pay per generated sentence; one click on “Generate” creates several. The price per sentence is on the button, and the total for the next click is in the line below it.

The Spark chat isn't visible at the moment: Spark starts straight with the style profile — the profile at the top, the fan context below it. Instead, you develop a style profile in a chat of your choice: under Style profile → “JSON guide” you copy the instructions and paste them there. You bring the finished profile back like this: “Info” tab → create a new info card → “Import from file” → choose the .json files; the profile lands in the Spark tab. For quick questions in between, there's the floating chat.

The { } icon at the top right opens the current Spark configuration (or, in the Styling tab, the emoji configuration) as JSON. “Copy” puts it on the clipboard, and then you paste it into a language model. The icon glows cyan until you've copied it once (copying is part of the Pro plan). Anyone can view it. Copying and saving as a file (“Copy”, “Save (.json)”) is part of the Pro plan (Agency too).

Emoji or Spark profiles (JSON) generated by a language model are imported via “Import from file” — a profile is created from them automatically. You choose the profile name yourself; the model doesn't make one up.