Presets

Below are presets I have made for various LLMs, for usage with SillyTavern. Any and all documentation will be within the presets themselves, some in the readme, other info inside comments.

TOON

A Gemini 2.5-pro preset designed primarily to write out cartoony stories! (But don't worry, it can be used to write serious stuff too!) This preset plays a little fast and loose, but also offers a large selection of toggles to help you steer the AI in the way you want. Or the opposite with a randomness toggle!
PRO — 1.3 1.2 1.1 1.0
FLASH — 1.0
Example Output
HTML Mode Example
Older HTML Mode Example
Example Output (Old) 1
Example Output (Old) 2

NOIR

For Gemini 2.5-pro, this preset makes the output read like it's a retelling by {{char}}, like a detective's monologue from some noir movie. Hence the name!
Has a tendency to do very long outputs when left unchecked, a paragraph limiter prompt is included.
1.1 1.0
Examples:
Hmm
Seizure
From personal experience, I can confirm this thought.
Example of how the notes prompt looks in effect. Using a different theme to differentiate from the codeblocks that one uses.

Old Presets:

These are all discontinued.
Claude Opus 3: 2.2 2.1 2.0 1.9 1.8 1.7 1.6
Claude Sonnet 3.0: 1.1 1.0
Simple version (for scenario/RPG cards, gives a lot of control to the in-card instructions): 1.3 1.2
ChatGPT4o (Based off an old version of KnottyJB): 1.1 1
Gemini 2.5 (Based on FluffPreset): 1.1 1.0

Regexes

As of writing, ST now allows you to link regexes with presets. Toon 1.3, and Toon FLASH come with regexes, so you will not need to download any of these if that's what you're using.
Regexes to hide thinking blocks:
Claude, GPT
Secondary CoT
GPT
GPT 2
Gemini
The secondary CoT regex is something for my older cards, before scoped regexes were added to ST. For the GPT ones, they hide the <del> tags, so if you've used Knotty previously and hidden the tag, you won't need these. Claude/GPT are different to Gemini because those were written before the whole reasoning model thing happened.