🔧 Upcoming Bots & Scripts Update — Version 9.0.0.100

Yes, another big update is coming.

I know. I know. At this point, my bots and scripts are starting to evolve like overfed RPG companions with too many skill trees.

But this update matters.

Over the past few months, I have learned a lot about how my bots actually behave in long chats, what people enjoy, what tends to break immersion, what works well with scripts, and what quietly causes problems after several replies.

I have also received outside inspiration and feedback from followers, users, testers, and fellow creators who helped me notice possible improvements in my own work. Sometimes a good suggestion, a bug report, a conversation, or simply seeing how someone else approaches bot creation is enough to unlock a better structure.

That feedback, mixed with my own testing, allowed me to refine my ideas, synthesize my workflow, clean up older systems, and automate some of the repetitive tasks that used to slow down bot creation and maintenance.

Until now, my bots and scripts had already reached their 8th major version. That is not nothing. Version 8 was already a huge step forward compared to my older bots.

But now, I am preparing to put Version 9.0.0.100 into production across my two accounts.

🌡️ About the Update Schedule

This update will take time.

There are many bots, many universes, many scripts, and many interconnected systems to update properly. I do not want to rush this and leave half-broken cards online.

Also, there is currently a heatwave, and I would rather not transform my apartment into a sauna by keeping my PC running all day.

So the update will happen progressively, universe by universe, or subgroup by subgroup.

Some bots may temporarily disappear from public access while I update them, clean them, test them, and reconnect them to the new script structure.

This is temporary maintenance, not a deletion wave.

The goal is not to erase your chats or remove your favorite bots. The goal is to bring them back cleaner, stronger, more stable, and more enjoyable to use.

I hope you will understand and forgive the temporary unavailability of some bots during this transition. It is a short-term inconvenience for a long-term improvement.

📌 What Is Happening to My Universes?

As previously announced, Concordia is now retired. I will no longer actively publish new Concordia bots.

The same will now happen with Afterlife. I still like the ideas behind it, but I have reached a point where my inspiration for that universe is already exhausted.

Some of my other active universes may also slow down for a while. If there are no new releases for some time, it probably means I have reached the end of my currently prepared bot stock for that setting.

This does not mean I am quitting. It simply means I would rather slow down than force weak bots just to keep a schedule alive.

I still accept requests and suggestions through my Discord, so if you have ideas, characters, scenarios, adaptations, or concepts you would like to see, feel free to come and propose them.

📚 Manuals Will Also Be Updated

The lore manuals and command manuals will also be revised soon.

This update changes several parts of my script architecture, and in some cases it also adjusts small details in my universes, bot structures, mechanics, commands, and runtime behavior.

Because of that, older manuals may become incomplete, outdated, or misleading. I will update them progressively so they match the new Version 9 systems.

🔒 Temporary Private Mode During Maintenance

When a universe or subgroup is being updated, I may temporarily privatize the affected bots.

This allows me to:

• update the bot card cleanly;

• replace older script references;

• fix outdated formatting;

• remove duplicated or obsolete instructions;

• connect the bot properly to the new Version 9 structure;

• test the bot before making it public again;

• avoid leaving half-broken bots online during maintenance.

Once updated, the bots will return public, cleaner and more polished.

This maintenance is not meant to wipe your histories. This is a bot-and-script update, not a deliberate reset of your chats.

⚙️ What Version 9.0.0.100 Actually Changes

This is not only a content update.

It is not just “more lore” or “more bots”.

It is mostly a deep structural update to how my bots use information, preserve coherence, respect their limits, react to the user, and remain playable over longer conversations.

In short: Version 9 turns my scripts into a cleaner, more stable, smarter RP architecture designed for long-form roleplay.

🧱 A Clearer Architecture

Before Version 9, my V7 and V8 scripts already gave a lot of information to the model: lore, rules, mechanics, reactions, adult context, save data, formatting reminders, command logic, and more.

But some of that structure was still too stacked.

Some scripts did several things at once. Some information was repeated. Some bots carried too much context directly inside their own files. And during long conversations, the model could eventually mix elements, forget important rules, or reuse context at the wrong moment.

With V9.0.0.100, each layer has a cleaner role.

• Core Runtime defines the shared logic of a universe or bot family.

• Main Lorebooks explain the world, canon, cultures, factions, species, and shared systems.

• Sublorebooks handle local context, specific groups, casts, regions, schools, branches, or scenario subsets.

• Scene Core maintains the current scene and keeps the roleplay situation playable.

• Gameplay manages active mechanics, progression, quests, routes, encounters, and scenario loops.

• Reaction Core improves how characters and NPCs respond to the situation.

• Adult Context Engine frames mature routes when they are relevant and supported.

• Post-Process acts as a final correction layer for format, tone, agency, and consistency.

The idea is no longer to throw a huge block of information at the model and hope it sorts everything correctly.

The idea is to give it the right information, in the right place, at the right time.

🧠 Smart Token Injection

One of the biggest improvements in this version is what I call Smart Token Injection.

The name sounds technical, but the idea is simple: the scripts are designed to manage more carefully what gets injected into the bot’s context.

Because in JanitorAI, every token matters.

If too much irrelevant lore is injected, the bot becomes heavy.

If the same rules are repeated everywhere, space is wasted.

If information is injected in the wrong place, the model may use it incorrectly.

And if everything stays active all the time, long conversations become less stable.

Smart Token Injection is meant to make context cleaner and more useful.

• General world information stays in the lore scripts.

• Local information stays in sublorebooks.

• Current scene elements stay in Scene Core.

• Active mechanics go into Gameplay.

• Character reactions go into Reaction Core.

• Mature routes stay inside Adult Context.

• Final corrections pass through Post-Process.

This reduces redundancy, avoids useless walls of text, and keeps more room for what is actually happening in the conversation.

In practice, this should create bots that hold context better, repeat themselves less, forget important rules less quickly, and remain more coherent after many messages.

This is less flashy than a new character, but for RP quality, it can change a lot.

🧹 Less Noise, More Useful Context

One of the biggest problems in older versions was that a bot could receive too much information at once.

Even when the information was good, it was not always useful for the current scene.

Version 9 tries to improve this by keeping stable elements in the right files, avoiding duplicates, and not multiplying tiny scripts when one clean module is enough.

This is why many deep context files, local scripts, and isolated mini-modules have been fused, moved, or reorganized.

If an important character belongs to a small universe or recurring cast, it is usually better to place that character directly in the relevant lorebook or sublorebook instead of loading a separate deep context script.

If several mechanics are strongly connected, it is usually better to unite them inside a clear Gameplay module instead of stacking three tiny scripts that risk stepping on each other.

The result should be easier for me to maintain, and more stable for you to use.

🎬 More Alive Scenes

Version 9 also improves how scenes move forward.

A bot should not always wait for {{user}} to do absolutely everything.

NPCs can keep talking, move around, work, argue, investigate, react to the environment, create complications, or push a situation forward.

But this does not mean the bot should decide for {{user}}.

This is one of the most important rules of the architecture: the world can move, but {{user}} stays free.

The bot may advance NPCs, dangers, locations, consequences, factions, and events.

But it should not control {{user}}’s thoughts, emotions, words, actions, body, consent, choices, or decisions.

That distinction is extremely important.

The goal is to make scenes more active without sacrificing player agency.

🎮 More Concrete Gameplay

Gameplay modules have also evolved a lot.

They are no longer just mood notes or vague reminders. They now support more concrete playable loops, depending on the bot and universe.

This can include:

• exploration;

• investigation;

• combat;

• survival;

• travel;

• relationship progression;

• party management;

• missions;

• confrontations;

• social events;

• horror escalation;

• dungeon crawling;

• dating sim routes;

• open-world roleplay;

• mature routes when the context supports them.

The bot should better understand what kind of scene is currently happening.

An investigation should leave room for clues, suspicion, false leads, and deduction.

A combat scene should preserve pressure, danger, positioning, and consequences.

An exploration scene should make places, risks, discoveries, and atmosphere matter.

A relationship scene should respect rhythm, tension, trust, attraction, distance, and choice.

A mature route should not swallow the whole scenario the moment it appears.

This is an important improvement against scenes that feel too flat, too automatic, or too eager to skip to the easiest outcome.

💬 More Credible Reactions

The Reaction Core becomes much more important in Version 9.

Its role is to help characters and NPCs react according to the actual context, not just according to generic chatbot habits.

It takes into account things like:

• the current situation;

• the relationship with {{user}};

• the danger level;

• the social status of {{user}};

• the groups present in the scene;

• previous actions;

• physical differences;

• the tone of the scene;

• the established consequences;

• the personality of the active characters.

The goal is to avoid generic reactions.

Less interchangeable NPCs.

Less automatic acceptance.

Less “everyone attacks for no reason”.

Less “everyone instantly loves the user”.

Less repetition of the same emotional beat forever.

More nuance, more context, more continuity, and stronger character identity.

🛡️ Stronger Guideline Handling

Another major part of this update is stronger structural guardrails.

Version 9 does not simply tell the model “please follow the rules”.

It organizes the scripts so they can better block, redirect, or correct elements that go against the bot’s guidelines, the project’s rules, or the platform’s expectations.

This concerns several things.

Respecting {{user}}’s Agency

The bot should not decide what {{user}} thinks, feels, says, accepts, refuses, does, suffers, or chooses.

It can create consequences, tension, danger, opportunities, reactions, and events.

But it should not take control of the user.

Respecting the Bot’s Promise

If a bot is built for a specific mood, canon, genre, or type of scene, the scripts help prevent drift that breaks that promise.

The bot should not mix unrelated contexts.

It should not invent mechanics that contradict its own setting.

It should not turn a serious scene into an out-of-character menu for no reason.

It should not force an adult route.

It should not resolve an entire plot too quickly just because that is easier.

Respecting Mature Limits

Adult content must remain framed by the bot’s context, relationship, tone, character ages, consent, and supported settings.

The Adult Context Engine is not there to push every scene toward sexual content.

It is there to handle mature content correctly when it is relevant, allowed, and appropriate for the bot.

Then the Post-Process layer reinforces those rules one last time.

This is not magic. No script can guarantee the model will never make mistakes.

But it is a major improvement: important rules are no longer just written somewhere. They are integrated into several layers of the architecture.

🧾 Better Long-Chat Continuity

Version 9 also improves scene memory and long-chat continuity.

The scripts are not trying to save the entire RP word for word. That would be too heavy and not very useful.

Instead, they focus on important compact facts:

• the current route;

• major choices;

• relationship dynamics;

• active dangers;

• useful flags;

• objectives;

• state changes;

• important past events;

• current scene pressure;

• ongoing consequences.

This is cleaner, lighter, and more useful than trying to remember everything.

The goal is for the bot to keep understanding what matters after 20, 40, or 80 messages.

This is often where older versions started to get tired. Version 9 is designed to make those long conversations more solid.

⌨️ Cleaner Commands

Version 9 also better respects the way I prefer commands to work.

The idea is to place a command at the beginning of a message, then continue directly in roleplay.

For example:

/command active

And the scene continues naturally in the same message.

This avoids sending isolated technical messages that pollute the chat history.

That matters because the model rereads previous messages. If the history is full of out-of-character command turns, the bot may become more likely to generate out-of-character replies.

When commands stay attached to roleplay messages, the conversation keeps a better rhythm.

🧩 Fewer Duplicates, Fewer Unnecessary Scripts

A lot of work in this update went into reducing redundancy.

When two scripts did almost the same thing, they were fused when it made sense.

When information was duplicated, it was moved back to the right place.

When a character or mechanic could be handled by an existing lorebook, I avoided adding an extra deep context script.

When a subgroup shared a foundation with another subgroup, I kept a common trunk instead of recreating a full stack for everything.

This makes the packs cleaner, easier to maintain, and less painful to install.

🚀 What Version 9 Improves Compared to Versions 7 and 8

Version 7 and Version 8 were already important steps, but they still had weaknesses.

Some systems were too scattered. Some scripts repeated similar instructions. Some bots carried too much information directly in their card when that information should have been handled by lorebooks or shared scripts.

Some command systems were not unified enough. Some gameplay modes were underdeveloped. Some bots needed better separation between character identity, world rules, scenario mechanics, and adult context.

Version 9 is meant to fix that.

Cleaner Structure

Personality files focus on character identity.

Scenario files focus on the situation, start state, routes, progression, mechanics, and runtime flow.

Lorebooks hold shared canon.

Scripts handle runtime behavior.

Adult systems stay in the adult layer.

Post-Process handles final response discipline.

This reduces confusion and makes bots easier to maintain.

Better Bot and Script Synergy

Version 9 is designed so bot files and scripts work together instead of repeating each other.

Older versions sometimes had the same rule written in the personality, scenario, lorebook, and script. That made bots heavier and harder to update.

Now, more information is placed where it belongs.

The bot card carries what is unique to the bot.

The universe script carries shared systems.

The sublorebook carries local canon.

The gameplay script carries interactive mechanics.

Stronger Character Behavior

Characters should not just answer. They should respond like themselves.

A proud character should not behave like a shy one.

A dangerous character should not become harmless without reason.

A loving character should not suddenly become cold because the bot forgot the relationship.

A chaotic character should still have a recognizable pattern instead of becoming random noise.

Version 9 helps preserve identity, tone, emotional logic, and relationship dynamics.

More Flexible Gameplay

Many of my newer bots are not simple chatbots anymore.

Some are adventures. Some are romance routes. Some are world simulators. Some are horror scenarios. Some are survival stories. Some are monster encounters. Some are interactive fiction systems with commands, HUD elements, progression, or optional mechanics.

Version 9 makes these gameplay styles easier to support without creating a completely different script stack for every bot.

Better Maintenance and Automation

This update also improves my own workflow.

By making templates stricter, structures cleaner, and script roles more predictable, I can update bots faster and with fewer mistakes.

This matters because I manage many bots, several universes, and two accounts.

The more consistent the structure becomes, the easier it is to create new bots, patch older ones, and keep entire universes compatible with the latest systems.

🧾 What Users Should Feel in RP

With V9.0.0.100, I hope users will experience:

• less passive bots;

• more alive scenes;

• fewer repetitions;

• fewer artificial menus;

• more credible NPCs;

• more playable mechanics;

• better long-chat continuity;

• less mixing between unrelated contexts;

• stronger respect for {{user}}’s agency;

• better handling of multi-character scenes;

• better adult framing when relevant;

• better correction of tone, canon, and guideline drift.

This is really a structure and stability update.

It is not only “more content”.

It is a better way to use that content.

🛠️ What I Will Be Updating

Depending on the bot, the update may include:

• Bio.html cleanup;

• personality restructuring;

• scenario restructuring;

• new or improved first messages;

• example dialog cleanup;

• official tag review;

• script assignment updates;

• lorebook connection updates;

• command manual updates;

• lore manual updates;

• runtime compatibility checks;

• minor universe corrections;

• better separation between public presentation and active mechanics.

Some bots may only need a light pass.

Others may need deeper surgery.

Older bots, especially those created before my current template standards, may change the most.

📦 In Summary

Versions 7 and 8 mostly gave the model a lot of material.

Version 9.0.0.100 gives it a smarter method for using that material.

Smart Token Injection helps inject less noise and more useful context.

The separation of layers makes scenes more coherent.

Gameplay becomes more concrete.

Reactions become more contextual.

Save logic becomes more compact.

Post-Process becomes more protective.

Guidelines are better integrated into the entire architecture.

The goal is simple: more stable, more alive, more respectful bots that can support real RP experiences over time.

💬 Requests Are Still Welcome

Even if some universes are slowing down or retiring, requests are still welcome on my Discord.

If you have a character idea, a scenario concept, a universe suggestion, a franchise adaptation, or a bot type you would like to see, you can suggest it there.

I cannot promise that every request will be made, but outside ideas have already helped me improve my work, and some of my favorite projects started because someone gave me the right spark at the right time.

❤️ Thank You

Thank you to everyone who uses my bots, follows my accounts, tests the systems, reports bugs, gives feedback, shares ideas, or simply enjoys the strange little worlds I keep building.

This update is big, and it will take time, but it is meant to make the whole experience better.

The biggest technical part is behind me now.

The most interesting part begins next: seeing how all of this behaves in actual play, with your characters, your choices, your scenes, your magnificent chaos, and your very personal ways of breaking my systems.

Thank you for your patience while I update everything.

And thank you for making this work feel worth continuing.

Version 9.0.0.100 is coming.

Slowly, carefully, probably while sweating near a fan, but it is coming.