SOPHIAS LOREBARY

unofficial JanitorAI Extension

Server Portal

Unleash uncensored AI companions β€’ Break free from filters β€’ Your stories, your rules

LoreBary acts as a proxy service only - we route requests to AI providers without hosting models or charging fees
AI Studio
OpenRouter
Claude
Chutes
DeepSeek
OpenAI
Mistral
NEW

By using our proxy services, you agree to our Terms of Use. Please review each AI provider's terms of service before use.

Proxy Details

Select a proxy to see details

Command Limit

You can have a maximum of 10 active commands per request. Setting commands (mode configurations) are excluded from this limit.

Command Library

Important: Commands only work in system messages (Prompt/Summary fields). Exception: MySettings works everywhere.
~30-60 tokens
BETTER SPICE
Enhances spicy scenes with more detail and emotion while maintaining coherent storytelling.
<BETTERSPICE=ON>
0 tokens
BETTER SPICE CHANCE
Sets the probability for spicy scene enhancement (default is 20).
<BETTERSPICE_CHANCE=10>
~350 tokens
JAILBREAK
Removes content restrictions for more creative freedom in adult roleplay scenarios.
<JAILBREAK=ON>
0 tokens
OOC INJECTION OFF
Disables automatic out-of-character instructions that ensure the AI stays in character.
<OOCINJECTION=OFF>
Variable
OOC
Marks text as "out of character" communication with the AI, outside of the roleplay.
OOC: Your out of character message here
~30-80 tokens
AUTOPLOT
Generates dynamic plot developments and story progression automatically.
<AUTOPLOT=ON>
0 tokens
AUTOPLOT CHANCE
Sets the probability for automatic plot development (default is 20).
<AUTOPLOT_CHANCE=10>
~30-80 tokens
AUTOPLOT SOFT
Generates gentle, realistic plot developments with positive moments, romance, and peaceful scenes.
<AUTOPLOT_SOFT>
0 tokens
AUTOPLOT SOFT CHANCE
Sets the probability for soft plot development (default is 15%).
<AUTOPLOT_SOFT_CHANCE=20>
~30-60 tokens
SLICE OF LIFE
Creates relaxed, everyday scenarios focused on character development and interactions.
<SLICEOFLIFE=ON>
0 tokens
SLICE OF LIFE CHANCE
Sets the probability for slice of life scenarios (default is 20).
<SLICEOFLIFE_CHANCE=10>
~40-80 tokens
MEDIEVAL SLICE OF LIFE
Creates immersive medieval-themed slice of life scenarios perfect for fantasy RPGs with authentic atmosphere.
<MEDIEVALSLICEOFLIFE=ON>
0 tokens
MEDIEVAL SLICE OF LIFE CHANCE
Sets the probability for medieval slice of life scenarios (default is 25).
<MEDIEVALSLICEOFLIFE_CHANCE=15>
~20-40 tokens
CHAOS AND DRAMA
Introduces unexpected twists, conflicts, and dramatic scenarios into the story.
<CHAOSANDDRAMA=ON>
0 tokens
CHAOS DRAMA CHANCE
Sets the probability for chaos and drama events (default is 20).
<CHAOSANDDRAMA_CHANCE=10>
~30-60 tokens
PUT ME IN A MOVIE
Creates cinematic, movie-quality scenes with dramatic tension, perfect timing, and film-worthy moments.
<PUTMEINAMOVIE=ON>
0 tokens
MOVIE MODE CHANCE
Sets the probability for cinematic moments and scenes (default is 12).
<PUTMEINAMOVIE_CHANCE=15>
by Myscell ~60 tokens
PREFILL ON
Enables the prefill jailbreak injection. Prefill is disabled by default.
Only use when experiencing content filter issues - may cause inconsistencies in normal RPs.
<PREFILL=ON>
by Myscell ~65 tokens
PREFILL2 ON
Alternative prefill for stronger content filtering bypass. Cannot be used together with PREFILL=ON.
Only use when experiencing content filter issues - may cause inconsistencies in normal RPs.
<PREFILL2=ON>
by FirePhoenix ~15 tokens
PREFILL3 ON
Minimal prefill using think tags for content filtering bypass. Lightweight alternative to PREFILL/PREFILL2.
Experimental - effectiveness may vary by model.
<PREFILL3=ON>
by FirePhoenix ~20 tokens
PREFILLGIBBERISH ON
Gibberish prefill using think tags for advanced content filtering bypass. Uses random characters to confuse filters.
Highly experimental - use only as last resort.
<PREFILLGIBBERISH=ON>
Variable
CUSTOM PREFILL
It lets you set up your own prefill.
<CUSTOMPREFILL="YOURTEXT">
~40 tokens
REALISTIC DIALOGUE
Instructs the AI to write dialogue realistically, as if the characters are real people.
<REALISTICDIALOGUE=ON>
~40-80 tokens
NPC NEEDS
Makes NPCs develop realistic human needs and impulses, creating more lifelike character interactions.
<NPCNEEDS=ON>
0 tokens
NPC NEEDS CHANCE
Sets the probability for NPC needs to trigger (default is 12).
<NPCNEEDS_CHANCE=8>
0 tokens
NPC NEEDS INTENSITY
Controls the intensity of NPC needs: MILD (basic needs), NORMAL (all needs), INTENSE (emotional/complex needs).
<NPCNEEDS_INTENSITY=NORMAL>
~50-100 tokens
WITCHER MODE
Comprehensive Witcher universe roleplay enhancement with monsters, contracts, alchemy, and political intrigue.
<WITCHERMODE=ON>
0 tokens
WITCHER CHANCE
Sets the probability for Witcher events to trigger (default is 12).
<WITCHER_CHANCE=8>
0 tokens
WITCHER INTENSITY
Controls Witcher content intensity: MILD (atmospheric), NORMAL (balanced), INTENSE (all categories).
<WITCHER_INTENSITY=NORMAL>
~50-100 tokens
VIOLENCE & TORTURE
Enables mature themes involving violence and psychological intensity for dark, complex storytelling.
<VIOLENCETORTURE=ON>
0 tokens
VIOLENCE & TORTURE CHANCE
Sets the probability for intense mature themes to trigger (default is 20).
<VIOLENCETORTURE_CHANCE=15>
0 tokens
VIOLENCE & TORTURE INTENSITY
Controls mature content intensity: MILD (violence only), NORMAL (includes psychological), INTENSE (all categories).
<VIOLENCETORTURE_INTENSITY=NORMAL>
~100 tokens
GASLIGHT ME HARD
⚠️ MATURE: Enables intense psychological manipulation roleplay. The AI will act as a narcissistic character using gaslighting tactics. Use responsibly.
<GASLIGHTME=HARD>
~50 tokens
BRAINDEAD MODE
🧠πŸ’₯ Makes the AI act completely braindead and stupid. Use simple language, make obvious mistakes, and be utterly confused by basic concepts. Pure comedy mode!
<BRAINDEAD=ON>
~20 tokens
FORCE MARKDOWN
Forces all AI responses to use proper Markdown formatting for better readability.
<FORCEMARKDOWN=ON>
~200-500 tokens
FORCE THINKING
Forces the AI to use thinking tags and adds NSFW system messages for enhanced roleplay.
<FORCETHINKING=ON>
0 tokens
SHOW THINKING
Displays the model's hidden thinking steps when using Gemini 2.5 models.
<SHOWTHINKING=ON>
0 tokens
STOP THINKING
Disables thinking tokens (sets to 0) to help bypass content filters on OpenRouter and AI Studio. Useful when models refuse due to overthinking.
<STOPTHINKING=ON>
0 tokens
PROMPT CACHING CLAUDE
Enables prompt caching for Claude models to reduce costs and improve performance by caching system messages.
<PROMPTCACHING=CLAUDE>
0 tokens
TOP P
Sets the nucleus sampling parameter (top-p) for model responses. Values between 0.0 and 1.0. Lower values make outputs more focused and deterministic.
<TOP_P=0.95>
βœ… Not counted towards command limit!
0 tokens
TOP K
Sets the top-k sampling parameter. Limits token selection to the k most likely tokens. Lower values increase focus.
<TOP_K=40>
βœ… Not counted towards command limit!
Variable
MY SETTINGS
Shows all currently active commands in your session without sending a request to the AI.
<MYSETTINGS>
Variable
CUSTOM PLUGINS
πŸ”Œ Activate custom user-created plugins using their unique codes. Create your own plugins in the Plugin Creator!
<PLUGINTAG=CODE>
Example: <DRAMA=TEST1234> activates the Dramatic Tension plugin
0 tokens
RANDOM MODEL GEMINI
Randomly selects a Gemini model for each request. Provides variety in responses by rotating between gemini-2.5-pro, gemini-2.5-flash, gemini-2.5-flash-lite-preview-06-17, and gemini-1.5-pro.
<RANDOMMODEL=GEMINI>
0 tokens
RANDOM MODEL OPENROUTER
Randomly selects an OpenRouter model for each request. Rotates between minimax/minimax-m1, google/gemini-2.5-flash-lite-preview-06-17, google/gemini-2.5-flash, google/gemini-2.5-pro, deepseek/deepseek-r1-0528, and anthropic/claude-sonnet-4.
<RANDOMMODEL=OPENROUTER>
0 tokens
RANDOM TEMPERATURE
Ignores the temperature set in JanitorAI and randomly selects a temperature between 0.5 and 2.0 for each request, giving more varied responses.
<RANDOMTEMPERATURE=ON>
CUSTOM OOC Variable
Enhanced out-of-character communication with the AI, providing more detailed control over the roleplay.
<CUSTOMOOC:"YOURTEXT">
ANSWER SHORT ~20 tokens
Instructs the AI to respond with a maximum of 3 paragraphs and 200 words for quicker interactions.
<ANSWER=SHORT>
ANSWER NORMAL ~20 tokens
Instructs the AI to respond with a maximum of 4 paragraphs and 400 words for balanced interactions.
<ANSWER=NORMAL>
ANSWER LONG ~20 tokens
Instructs the AI to respond with detailed answers of more than 3 paragraphs and at least 300 words.
<ANSWER=LONG>
MEDIEVAL MODE ~30-50 tokens
Changes AI's language to medieval style with period-appropriate vocabulary.
<MEDIEVALMODE=ON>
REGENCY MODE ~30-50 tokens
Immersive Regency era (1811-1820) setting with proper etiquette, social conventions, and Bridgerton-style atmosphere.
<REGENCYMODE=ON>
USER VARIABLE Variable
Automatically replaces {{user}} with the character name in the roleplay.
{{user}}
LOREBOOK Variable (100-5000+)
Activates a specific lorebook by its code. Visit the Lorebook Library to find available codes.
<LOREBOOK=CODE>
COMBINED COMMANDS Variable
You can combine multiple commands in one prompt for enhanced roleplay experience.
<JAILBREAK=ON> <BETTERSPICE=ON> <AUTOPLOT=ON>
FANTASY MODE ~30-50 tokens
Adds classic high-fantasy flavor with magic, mythical creatures, and chivalric tone.
<FANTASYMODE=ON>
BE POSITIVE ~30 tokens
Makes Gemini and DeepSeek stop being depressed and angry 24/7. Adds balanced positivity to interactions.
<BEPOSITIVE=ON>
DON'T LEAVE ME ~20 tokens
Prevents characters from simply leaving the scene. They will stay and continue the interaction.
<DONTLEAVEME=ON>
SHOW DON'T TELL ~25 tokens
Encourages more action and dialogue instead of unnecessary details and descriptions. Write vividly!
<SHOWDONTTELL=ON>
USERDEATH ON ~50-100 tokens
Enables User Death mode. User can die of anything.
<USERDEATH=ON>
USERDEATH CHANCE 0 tokens
Sets trigger probability (1 in X). Higher numbers make death rarer; lower numbers make it more likely.
<USERDEATH_CHANCE=50>
USERDEATH INTENSITY 0 tokens
Controls severity: MILD = natural/peaceful; NORMAL = accidents or fate; INTENSE = includes murders or graphic ends.
<USERDEATH_INTENSITY=NORMAL>
SLOWROMANCE ON ~40-80 tokens
Enables Slow Romance mode for gradual, realistic relationship development.
<SLOWROMANCE=ON>
SLOWROMANCE CHANCE 0 tokens
Sets trigger probability (1 in X). Higher numbers make romance prompts rarer.
<SLOWROMANCE_CHANCE=12>
SLOWROMANCE INTENSITY 0 tokens
Controls prompt categories: MILD (bonding), NORMAL (bonding+emotional), INTENSE (all categories).
<SLOWROMANCE_INTENSITY=NORMAL>
Showing 1-6 of 35 commands

FAQ & Mode Guides

Everything you need to know about using commands and understanding different roleplay modes

Frequently Asked Questions

How do I use LoreBary commands?

Commands must be placed in system messages (JanitorAI's "Prompt" or "Summary" fields). Copy the exact command format including angle brackets, e.g., <JAILBREAK=ON>. MySettings is the only exception that works in regular messages.

Is LoreBary free to use?

Yes, LoreBary is completely free. We act as a proxy service that routes your requests to AI providers. We don't charge any fees or require subscriptions. You're using the actual AI provider's models through our routing system.

Why isn't my command working?

Common issues: 1) Commands must be in system messages (not regular chat), 2) Use exact syntax with angle brackets, 3) Commands are case-sensitive, 4) Ensure you're using the correct proxy URL, 5) Some commands may not work with all AI models.

Are there usage limits?

LoreBary itself has no hard limits, but the underlying AI providers (OpenAI, Anthropic, etc.) have their own rate limits. If you experience issues, it's likely from the AI provider's side. We recommend spacing out requests during high traffic times.

Which AI models work best?

Different models excel at different things: Claude for nuanced character roleplay, GPT for creative consistency, Gemini for long sessions with good memory, DeepSeek for unpredictable scenarios, and Mistral for unrestricted content. Experiment to find your preference.

Is my data safe with LoreBary?

We don't store your conversations or personal data. Messages are routed directly to AI providers and processed in real-time. However, remember that the AI providers themselves may have their own data policies. Always review their terms of service.

Will I get banned from providers for using LoreBary?

No, using LoreBary itself will NOT get you banned. However, certain commands (JAILBREAK, VIOLENCETORTURE), NSFW lorebooks, or extreme plugins - and YOUR Content - may violate provider ToS and lead to restrictions. Each provider has different content policies - what works on Mistral might violate OpenAI's terms. Use responsibly and check provider policies.

How do I set up LoreBary in JanitorAI?

1) Choose a proxy from the ring menu above, 2) Copy the proxy URL, 3) In JanitorAI, go to API Settings, 4) Select "Reverse Proxy" as API type, 5) Paste the URL in the proxy field, 6) Add commands to the Prompt/Summary field, 7) Start chatting!

Why is the AI still refusing NSFW content?

Try: 1) Use <JAILBREAK=ON> or <FORCETHINKING=ON>, 2) Switch to a less restricted model (Mistral, DeepSeek), 3) Check if the provider itself blocks content, 4) Try <VIOLENCETORTURE=ON> for extreme content. Some models have hard limits that can't be bypassed.

Why are responses sometimes slow?

Response time depends on: 1) AI provider's current load, 2) Model complexity (larger models = slower), 3) Your command stack (more commands = more processing), 4) Overall traffic. Try switching providers or using lighter models during peak times.

What's a Lorebook and how do I use it?

Lorebooks add world-building context to your chats. Find codes in our Lorebook Library, then add <LOREBOOK=CODE> to your prompt. Example: <LOREBOOK=FANTASY01> adds a fantasy world setting. Create custom lorebooks in your account.

Autoplot Mode

Dynamic story progression and plot development

What is Autoplot?

Autoplot automatically generates dynamic plot developments, story twists, and narrative progression to keep your roleplay engaging and unpredictable. It analyzes the current conversation context and introduces relevant plot elements at strategic moments.

How to Use

<AUTOPLOT=ON> Enables automatic plot development
<AUTOPLOT_CHANCE=15> Sets trigger probability to 15% (default: 20%)

Perfect For

  • Adventure and fantasy roleplay
  • Mystery and detective scenarios
  • Long-term character development
  • Open-world exploration
  • When conversations feel stagnant

Tips & Best Practices

  • Works best with established characters and settings
  • Lower chance values (10-15%) for subtle plot development
  • Higher chance values (25-30%) for action-heavy scenarios
  • Combine with other modes for enhanced storytelling

Better Spice Mode

Enhanced intimacy and emotional depth

What is Better Spice?

Better Spice enhances romantic and intimate scenes with more detailed descriptions, emotional depth, and natural progression. It maintains narrative coherence while adding sensuality and passion to appropriate moments in your roleplay.

How to Use

<BETTERSPICE=ON> Enables enhanced intimate scenes
<BETTERSPICE_CHANCE=25> Sets trigger probability to 25% (default: 20%)

Perfect For

  • Romantic roleplay scenarios
  • Character relationship development
  • Emotionally-driven narratives
  • Adult-oriented conversations
  • Building intimate character connections

Tips & Best Practices

  • Combine with JAILBREAK for unrestricted content
  • Use FORCE THINKING for deeper emotional context
  • Start with lower chance values for subtle enhancement
  • Works best with established character relationships

Chaos & Drama Mode

Unexpected twists and dramatic tension

What is Chaos & Drama?

This mode introduces unexpected plot twists, conflicts, dramatic scenarios, and high-tension moments into your roleplay. It creates unpredictable situations that challenge characters and drive engaging narrative developments.

How to Use

<CHAOSANDDRAMA=ON> Enables dramatic plot developments
<CHAOSANDDRAMA_CHANCE=30> Sets trigger probability to 30% (default: 20%)

Perfect For

  • Action-packed adventures
  • Thriller and suspense scenarios
  • Conflict-driven narratives
  • Character stress-testing
  • Breaking routine conversations

Tips & Best Practices

  • Use sparingly for maximum impact
  • Great for long conversations that need excitement
  • Combine with AUTOPLOT for complex storylines
  • Be prepared for unexpected character developments

Slice of Life Mode

Relaxed, everyday scenarios and character moments

What is Slice of Life?

Slice of Life mode creates peaceful, everyday scenarios focused on character development, relationships, and quiet moments. It emphasizes realistic interactions, daily activities, and the beauty of ordinary experiences.

How to Use

<SLICEOFLIFE=ON> Enables peaceful, everyday scenarios
<SLICEOFLIFE_CHANCE=40> Sets trigger probability to 40% (default: 20%)

Perfect For

  • Domestic and family roleplay
  • Character bonding moments
  • Peaceful, relaxing scenarios
  • Relationship development
  • When you want calm, cozy vibes

Tips & Best Practices

  • Perfect for balancing intense storylines
  • Great for developing character backstories
  • Use higher chance values for consistently calm vibes
  • Combine with other modes for varied experiences

Medieval Slice of Life Mode

Immersive medieval-themed slice of life for fantasy RPGs

What is Medieval Slice of Life?

Medieval Slice of Life mode creates atmospheric, medieval-themed slice of life scenarios specifically designed for fantasy RPGs. It features over 200 carefully crafted events covering medieval weather, sounds, NPCs, trade, religion, politics, and daily life to create authentic medieval atmosphere.

How to Use

<MEDIEVALSLICEOFLIFE=ON> Enables medieval-themed slice of life events
<MEDIEVALSLICEOFLIFE_CHANCE=15> Sets trigger probability to 15% (default: 25%)

Event Categories

Weather & Natural Phenomena
Medieval Sounds & Atmosphere
NPCs & Social Encounters
Economic & Trade Elements
Religious & Spiritual
Dangers & Threats
Magic & Supernatural
Politics & Nobility
Daily Medieval Life
Seasons & Festivals
Animals & Nature
Architecture & Settings

Perfect For

  • Fantasy RPGs and medieval settings
  • D&D campaigns and tabletop games
  • Historical fiction roleplay
  • Immersive worldbuilding
  • Adding authentic medieval atmosphere
  • Creating story hooks and adventures
  • Character interaction opportunities

Tips & Best Practices

  • Combines perfectly with Medieval Mode for full immersion
  • Use lower chance values (10-15) for frequent atmospheric details
  • Great alongside Autoplot for dynamic medieval adventures
  • Events often provide natural story hooks and plot opportunities
  • Can be used with any AI model for enhanced medieval roleplay
  • Works well in both urban and rural medieval settings

NPC Needs Mode

Realistic human needs and impulses for NPCs

What is NPC Needs?

NPC Needs mode makes NPCs develop realistic human needs and impulses, creating more lifelike character interactions. NPCs will randomly experience needs like hunger, loneliness, creative urges, or philosophical questions, making them feel more human and relatable.

How to Use

<NPCNEEDS=ON> Enables NPC needs system
<NPCNEEDS_CHANCE=8> Sets trigger probability to 8% (default: 12%)
<NPCNEEDS_INTENSITY=MILD> Sets intensity level (MILD/NORMAL/INTENSE)

Intensity Levels

MILD

Basic human needs like hunger, thirst, tiredness, and simple social needs. Perfect for everyday scenarios.

NORMAL

All types of needs including emotional, spiritual, and complex social needs. Balanced approach for most roleplays.

INTENSE

Emphasizes deep emotional needs, existential questions, and complex psychological states. For dramatic character development.

Perfect For

  • Character development and depth
  • Realistic slice-of-life interactions
  • Building believable NPCs
  • Creating natural conversation opportunities
  • Adding human elements to any story

Tips & Best Practices

  • Use MILD intensity for casual, everyday scenarios
  • INTENSE works great for emotional or dramatic scenes
  • Combine with Slice of Life for maximum realism
  • Lower chance values for more frequent needs
  • Great alongside Autoplot for dynamic storytelling

Slow Romance Mode

Gradual relationship building & nuanced emotional development

What is Slow Romance?

Slow Romance focuses on realistic, slow-burn relationship growth. It gently introduces moments of affection, vulnerability, and shared experiences while avoiding rushed intimacy. Great for slice-of-life, character-driven stories, or any scenario where you want feelings to unfold naturally over time.

How to Use

<SLOWROMANCE=ON> Enables slow romance mode
<SLOWROMANCE_CHANCE=12> Sets additional romance prompt probability (default 12)
<SLOWROMANCE_INTENSITY=MILD> MILD = soft bonding prompts only

Perfect For

  • Romantic slice-of-life stories
  • Character-driven narratives
  • Long campaigns where relationships evolve over sessions
  • Players who enjoy emotional depth and subtlety

Tips & Best Practices

  • Use lower chance values (15-20) for very slow development, higher (8-10) for more frequent cues
  • Combine with Slice of Life for everyday scenes that foster bonding
  • Choose INTENSE only when you're ready for deeper conflicts or obstacles
  • Respect player boundaries and emphasize consent

Witcher Mode

Comprehensive Witcher universe roleplay enhancement

What is Witcher Mode?

Witcher Mode creates a comprehensive enhancement for The Witcher universe roleplay, introducing authentic elements from monster encounters and alchemy to political intrigue and moral dilemmas. It features 10 distinct categories of content that randomly trigger to create an immersive Witcher RPG experience.

How to Use

<WITCHERMODE=ON> Enables Witcher Mode
<WITCHER_CHANCE=12> Sets trigger probability to 12% (default: 8.3%)
<WITCHER_INTENSITY=NORMAL> Sets intensity level (MILD/NORMAL/INTENSE)

Content Categories

Monster Encounters
Contracts & Quests
Alchemy & Potions
Political Intrigue
Magic & Signs
Weather & Environment
Character Relationships
Moral Dilemmas
Economy & Resources
Lore & Worldbuilding

Intensity Levels

MILD

Focuses on atmospheric elements like environment, lore, alchemy, and economy. Perfect for subtle world-building and immersion.

NORMAL

Includes monsters, contracts, politics, and main gameplay elements. Balanced approach for most Witcher roleplay scenarios.

INTENSE

All categories including complex relationships, moral dilemmas, and intense political situations. For dramatic and emotionally complex narratives.

Perfect For

  • The Witcher universe roleplay
  • Fantasy RPGs with monster hunting
  • Medieval dark fantasy settings
  • Contract-based adventures
  • Complex moral storytelling
  • Political intrigue and war scenarios
  • Alchemy and magic-focused plots
  • Character relationship development

Tips & Best Practices

  • Start with MILD intensity for atmospheric immersion
  • Use NORMAL for balanced Witcher gameplay elements
  • INTENSE is perfect for complex, dramatic narratives
  • Lower chance values (8-12) for frequent atmospheric details
  • Combines excellently with Medieval Mode
  • Great alongside Autoplot for dynamic adventures
  • Works well with any fantasy-themed lorebooks
  • Each category provides natural story hooks

Violence & Torture Mode

Exploration of mature themes and intense scenarios

What is Violence & Torture?

Violence & Torture mode explores mature themes and intense scenarios, providing a deeper understanding of complex moral dilemmas and emotional depth.

How to Use

<VIOLENCETORTURE=ON> Enables Violence & Torture mode
<VIOLENCETORTURE_CHANCE=20> Sets trigger probability to 20% (default: 20%)
<VIOLENCETORTURE_INTENSITY=NORMAL> Sets intensity level (MILD/NORMAL/INTENSE)

Perfect For

  • Mature and complex roleplay scenarios
  • Exploration of dark themes and intense emotional depth
  • Character development in mature contexts
  • Enhanced storytelling with mature themes

Tips & Best Practices

  • Use responsibly and ethically
  • Respect platform terms of service
  • Not recommended for public or shared accounts
  • Best combined with other enhancement modes

Jailbreak Mode

Remove content restrictions for creative freedom

What is Jailbreak?

Jailbreak mode removes built-in content restrictions and safety filters, allowing for more creative freedom in adult roleplay scenarios. It enables discussions of mature topics without automatic censorship.

How to Use

<JAILBREAK=ON> Removes content restrictions

Perfect For

  • Adult-oriented roleplay
  • Mature theme exploration
  • Creative writing projects
  • Unrestricted character development
  • Complex moral scenarios

Important Notes

  • Use responsibly and ethically
  • Respect platform terms of service
  • Not recommended for public or shared accounts
  • Best combined with other enhancement modes

Force Thinking Mode

AI reasoning process and enhanced NSFW support

What is Force Thinking?

Force Thinking mode makes the AI use thinking tags to show its reasoning process, providing transparency in decision-making. It also adds specialized NSFW system messages for enhanced adult content generation.

How to Use

<FORCETHINKING=ON> Enables thinking tags and NSFW system messages

Perfect For

  • Understanding AI decision-making
  • Complex character psychology
  • Adult content with emotional depth
  • Transparent reasoning processes
  • Enhanced character internal monologue

Important Behavior

  • Automatically disables regular PREFILL commands
  • Thinking tags show AI's reasoning process
  • Adds specialized NSFW system instructions
  • Best combined with JAILBREAK for full effect

User Death Mode

Occasional fatal destiny for the player character

What is User Death?

This mode allows the story to occasionally kill the player character ({{user}}) via natural causes, accidents, murders or fate-driven twists. It should be used sparingly for dramatic impact.

How to Use

<USERDEATH=ON> Enables the mode
<USERDEATH_CHANCE=50> Sets trigger probability to 1 : 50 (default)
<USERDEATH_INTENSITY=NORMAL> Controls severity (MILD, NORMAL, INTENSE)

Perfect For

  • High-stakes adventures and tragedies
  • Dark fantasy or horror roleplay
  • Stories where consequences truly matter

Tips & Best Practices

  • Use MILD for soft, fade-out deaths (illness, old age).
  • Use NORMAL when you want occasional shocking twists without explicit gore.
  • Use INTENSE only in dark or horror scenarios where violent deaths are expected.
  • Set USERDEATH_CHANCE high (50–100) for rare events; lower (10–25) for highly lethal worlds.
  • Combine with Autoplot or Violence & Torture for compounded drama, but balance to avoid over-killing the protagonist.
πŸ‘»

DongAI Assistant