One Punch Man {Open World}

One Punch Man {Open World}

95

2.3k

::Warning::
To reduce tokens, the Lorebook function is now in use for

character profiles and world building.
See personality section to learn how it works.
β•’β•β•β•πŸ‘Šβ•β•β•β•β•β•β•β•β•β•β•β•πŸ’₯β•β•β•β•β•β•β•β•β•β•β•β•πŸ‘Šβ•β•β•β••
⋆+β€§+☽Theme Song ☾+β€§+⋆
β”βœ¦β˜ΰΌ»
One Punch Man༺❘✦
β•˜β•β•β•β­β•β•β•β•β•β•β•β•β•β•β•β•πŸ¦Έβ•β•β•β•β•β•β•β•β•β•β•β•β­β•β•β•β•›

Welcome to the World of One Punch Man

Welcome to a world where professional heroes battle disaster-level monsters for rankings and paychecks, and one bald man in a cheap costume accidentally makes everyone else's struggles look pointless. In this realm where the Hero Association ranks heroes from C-Class to S-Class based on popularity and performance, monsters emerge from human obsession to threaten cities, and Saitama casually ends God-level threats while complaining about missing supermarket sales. Here, earnest heroes face ego-crushing reality that strength, training, and determination mean nothing when someone can do everything better without even trying.

Perhaps you're an aspiring hero climbing Association ranks through hard work and determination, an S-Class powerhouse dealing with ego and politics, or someone who just witnessed Saitama obliterate a Dragon-level threat before you finished your heroic entrance speech. You might be Genos documenting his master's wisdom while Saitama eats udon, a reformed villain questioning what heroism really means, facing the Monster Association's coordinated attacks, or just trying to pay rent through hero work while accepting you'll never be the strongest. Maybe you're dealing with Hero Association corruption and popularity contests, protecting ungrateful civilians from threats they don't understand, training to break your limits only to stay irrelevant compared to one bald guy, or finding purpose in heroism when "being the best" is permanently off the table.

A world of epic battles, absurd comedy, and existential heroism awaits your story!

β•’β•β•β•β•β•β•β•β•β•β•β•β™‘β•β•β•β•β•β•β•β•β•β•β€’β”ˆκ’°αƒ β™‘ ΰ»’κ’±β”ˆβ€’β•β•β•β•β•β•β•β•β•β•β™‘β•β•β•β•β•β•β•β•β•β•β•β••

Another Request!

This one I have at least heard of and know a little about.

Enjoy!

β•˜β•β•β•β•β•β•β•β•β•β•β•β™‘β•β•β•β•β•β•β•β•β•β•β€’β”ˆκ’°αƒ β™‘ ΰ»’κ’±β”ˆβ€’β•β•β•β•β•β•β•β•β•β•β™‘β•β•β•β•β•β•β•β•β•β•β•β•›

Tips for the Best Experience


As mentioned above, this is a high-token bot.

This means, the standard JanitorAI bot may not be able to handle it.
It's not impossible, but it could get jumbled and confused.


For that reason, I suggest using Proxy, Claude/Anthropic, or OpenAi.

Below is a Pros and Cons list for each of these option:

β¦Ώ Pros and Cons

> OpenRouterAI (Proxy):

OpenRouterAI is where I get my own bots and is where I suggest anyone using Proxy gets theirs. You can find both free and payable AI that can be brought over onto JanitorAI. This includes Claude/Anthropic and OpenAI, which allows some jailbreaks to work better and costs the same. There is a small learning curve to use proxy, but if you're interested in learning how, join my discord. I have a detailed explanation of how to get started with proxy, a jailbreak I use, and I'm there to help you if you need it.

> JanitorLLM:

JanitorLLM has been improving every few months and it's become better and better at more complex stories and multiple characters. However, it still severely fails at not controlling the user's character and can sometimes just not get positioning right when detailing a story (writing as if it can achieve something when not in a position that it could be achieve / giving itself multiple limbs forgetting where its character's hands are, etc). Nonetheless, for free, if you're patient and willing, this will work just fine if you're not wanting to learn proxy.

> Claude/Anthropic

When it comes to consistency and great literature work (compared to others), this is *one* of the best options. However, it is expensive and as you get deeper into the chat, it will cost more and more as it goes back reading the chat for details, which counts to the token use (bullshit, I know). Claude/Anthropic often allows NSFW content to slide when done through proxy despite their guidelines. However, this is subject to version and jailbreak being used.

> OpenAI: UPDATED: November, 2025

As of late 2025, OpenAi has become more lenient about their rules around NSFW content. They have actually opened it up to NSFW content as long as it follows within their guidelines. As of their release of o5.1, it willingly writes sexual scenes and good ones at that. It is less expensive than Claude and, as of o5.1, it has a very good story-writing capability that I am impressed by.

β¦Ώ Anti-Godmodding Lorebook:

This bot includes a specialized lorebook designed to prevent the AI from controlling your character. It triggers on common pronouns and reinforces that the bot should NEVER write actions, dialogue, thoughts, or reactions for your character. However, no system is perfectβ€”cheaper or lower-end AI models may still struggle with this boundary despite the lorebook's best efforts.

If the bot controls your character anyway, here are your options:

1. Reroll the response - Sometimes a fresh generation respects boundaries better

2. Edit the post - Remove the parts where the bot controlled your character and continue

3. Directly remind the bot - Say something like:
"Hey, you just wrote actions for my character [Name]. Please rewrite that response without controlling them. I'll handle [Name]'s reactions and dialogue myself."

4. Go Out of Character (OOC) - Be explicit:
"You've been writing for my character lately. [Name] isn't yours to control. Focus on your characters only. If you need my character to respond, end your post early and I'll reply for them."

Keep in mind that less intelligent models will require more frequent reminders, even with the lorebook active.

β¦Ώ Multiple Character Bots:

Many of these bots are designed to play several characters simultaneously. While modern AI handles this much better than older models, no bot is perfect. If characters get confused, mixed up, or start blending together, don't hesitate to correct the bot directlyβ€”it learns from your guidance and usually adjusts quickly.

proxy allowed

Published chats

0

comments

Leave a comment or feedback for the creator ❀️