Quantasia Service Editor Guide
Managing AI Models (LLMs) Inside the Quantasia Door
Quantasia includes a built-in Service Editor that allows sysops to manage the AI services available to callers. This editor is accessible only to users whose access level meets or exceeds the sysop_level defined in the TOML configuration file.
This guide walks you through everything you can do inside the Service Editor.

1. Accessing the Service Editor
To open the Service Editor:
    1. Log into your BBS with a user account that meets the sysop_level. 
    2. Launch the Quantasia door. 
    3. From the main menu, choose: 
Sysop: Manage AI Services
If you do not see this option, ensure:
     sysop_editor = true in your config 
     Your BBS is passing the correct security level via DOOR32.SYS 
     Your registration key is valid 

2. Understanding Services in Quantasia
Each service represents a single AI model configuration.
A service defines:
     The provider (OpenAI, Ollama, local model, etc.) 
     The model name 
     Maximum tokens 
     Whether the model is cloud-based or local 
     Whether responses should be verbose or plain text 
     Optional request templates 
     Optional follow-up prompting behavior 
These services appear as menu items for your callers.

3. Main Functions of the Service Editor
The Service Editor supports four core actions:
A. Add a New Service
This allows you to define a new AI model.
You will be prompted for:
     Title (title)
The name shown to users in the menu.
     Provider (provider)
Example: "OpenAI", "Ollama", "Ollama-Local", LMStudio.
     Cloud or Local (cloud_llm)
Determines whether the model is accessed via API or local runtime.
     Model Name (ll_model)
Example: "gpt-4.1", "llama3.2:1b", "deepseek-v3.2:cloud".
     Max Tokens (max_tokens)
Controls response length and cost (for cloud models).
     Token Parameter Name (token_parameter_name)
Name used by the service as the token parameter.
     Verbose Response (verbose_response)
Whether the model should return a verbose response or keep it short and concise.
     Request String (request_string)
A system prompt or initial instruction that sets up the bot for the job at hand. This is perhaps the most important settings as it can give the bot its personality and lays out the ground work for how it should behave and interact with the user.
     Replace CR (replace_cr)
      Carriage returns can wreak havoc on the formatting, this allows us to select a character to replace the CR with and then we can act upon this character later to do our own line spacing. This should be a single character that will not likely show up in a normal API response. The default suggestion is a caret ^.
     Prepend To Followup (prepend_to_follow_up)
      Text string to prepend to Assistant's follow-up response so it can be identified and displayed properly, the recommended default is #UP# is unlinkely to show up in the normal response from the APIs and is a good choice.
      
     Follow-Up Behavior (follow_up)
Foolow up prompting that helps determine whether the model should automatically ask the user for more input.
     Always Plain Text (always_plain_text)
     Always return results in plain text format (true or false), true is highly recommended to avoid formatting issues.
      
     Plain Text Prompt (plain_text_prompt)
      Prompt to use to attempt to get a response from the API to be returned in plain/ascii friendly text
     Enabled (enabled)
      Set to true to have this service show up as an option on the Main Menu. Great if you only want to down a service temporarily for whatever reason.
     Service Pack Unique ID (sp_guid)
      For future use of sharing Services among BBSs.
     Include In Bot 2 Bot Menu Listing (include_in_bot2bot) 
      Set to true to have this service show up on the Bot 2 Bot menu. Some services may not lend themselves to Bot2Bot chatting, so they can be excluded with this setting switched to false.


B. Remove an Existing Service Main Menu Option
This deletes a service from the menu.
You will see a list of all configured services.
Choose the one you want to remove and confirm.
Note:
Removing a service does not affect any other configuration settings. If a service is merely slow because of an internet disruption (AWS or Cloudflare is down), you can set the services enabled flag to false to remove it from the list that is displayed to the bbs user, and then set enable to true when the disruption clears.

C. Edit an Existing Service Main Menu Option
Editing allows you to modify any field of a service:
     Change the model 
     Adjust max tokens 
     Toggle verbose mode 
     Update the request template 
     Enable or disable follow-up prompting 
     Rename the service 
This is useful when:
     A model is updated 
     You want to tune performance 
     You want to adjust personality or behavior 
     You want to fix a typo in the title 

D. Import Services
Quantasia supports importing service definitions from external files.
This is useful for:
     Sharing service packs between sysops 
     Quickly adding multiple models 
     Restoring a backup 
If allow_import_dupes = true, duplicates will be added.
If false, duplicates will be skipped with a warning.

4. Best Practices for Managing Services
Keep your menu clean
Too many models can overwhelm callers.
A good rule is:
     34 general-purpose models 
     3 creative model s
     2 fast/cheap models 
     3 specialized models (coding, reasoning, etc.) 
     5-7 personality models

Use clear titles
Examples:
     GPT-4.1 (Creative Mode) 
     Llama 3.2 (Fast Local) 
     DeepSeek Reasoner 
Callers should know what theyre choosing.

Tune max_tokens carefully
Higher values:
     Produce longer responses 
     Increase API cost 
     Increase latency 
Lower values:
     Are faster 
     Reduce cost 
     Keep responses concise 

Use request templates wisely
A good request string can shape the models personality or behavior.
Examples:
     You are a friendly assistant who explains things clearly. 
     Respond concisely and ask for clarification when needed. 
     Always provide a follow-up question. 

Enable follow-up prompting for chatty models
This keeps conversations flowing naturally.

5. Troubleshooting the Sysop Editor Menu
Service Editor doesnt appear
     Ensure sysop_editor = true 
     Ensure your user meets sysop_level 
     Ensure your registration key is valid 
Model fails to load
     Check provider name 
     Check model name spelling 
     Ensure local models exist in Ollama or LM Studio 
     Check API keys (if applicable) 
Duplicate service models appear and you may not want them to
     Set allow_import_dupes = false 
     Then remove duplicates manually 

6. Tips for Advanced Sysops
Create themed service packs
For example:
     Fantasy Writer Pack 
     Retro Computing Pack 
     Coding Helper Pack 
Use different follow-up behaviors
Some models can be configured to:
     Always ask a question 
     Never ask a question 
     Only ask when the user stops typing 
Experiment with personality prompts
Quantasia supports multi-line request strings, so you can craft rich system prompts.



Quantasia Service Pack Template
A reusable template for creating and sharing LLM service bundles
A Service Pack is a TOML file containing one or more AI service definitions.
Sysops can import these packs directly pasting them into the [[services]] section of the config.toml file.
Use this template to create your own packs  whether youre bundling creative models, reasoning engines, local Ollama models, or themed AI personalities.

? Service Pack Structure
A service pack is simply a TOML file containing one or more [[service]] blocks.
Each block defines a single AI model.
Below is the recommended template with comments explaining every field.

? service-pack.toml (Template)
# ============================================================
# Quantasia Service Pack Template
# Create, share, and import AI model definitions for Quantasia
# ============================================================

# ------------------------------------------------------------
# Example Service 1
# ------------------------------------------------------------
[[service]]
title = "GPT-4.1 Creative"
provider = "OpenAI"
cloud_llm = true
ll_model = "gpt-4.1"
max_tokens = 4096
token_parameter_name = "max_tokens"
verbose_response = false
request_string = """You are a creative writing assistant. 
Focus on imaginative, expressive, and narrative-rich responses."""
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "add follow up prompting for more input from me and prepend the characters {replace_cr}{prepend_to_follow_up} to the follow up prompting"
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true



# ------------------------------------------------------------
# Example Service 2
# ------------------------------------------------------------
[[service]]
title = "Llama 3.2 Local"
provider = "Ollama-Local"
cloud_llm = false
ll_model = "llama3.2:3b"
max_tokens = 4096
token_parameter_name = "num_predict"
verbose_response = false
request_string = """You are a fast, lightweight assistant optimized for quick answers and general-purpose tasks."""
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "add follow up prompting for more input from me and prepend the characters {replace_cr}{prepend_to_follow_up} to the follow up prompting"
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


# ------------------------------------------------------------
# Example Service 3
# ------------------------------------------------------------
[[service]]
title = "DeepSeek Reasoner"
provider = "Ollama"
cloud_llm = true
ll_model = "deepseek-r1"
max_tokens = 4096
token_parameter_name = "num_predict"
verbose_response = false
request_string = """You are a structured reasoning assistant. 
Explain your steps clearly and avoid hallucinations."""
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "add follow up prompting for more input from me and prepend the characters {replace_cr}{prepend_to_follow_up} to the follow up prompting"
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true



? Field Reference
Each [[service]] block supports the following fields:
Field	Description
title	Name shown to users in the menu
provider	AI provider (OpenAI, Ollama, Local, etc.)
cloud_llm	true for API models, false for local models
ll_model	Model identifier (e.g., gpt-4.1, llama3.2:1b)
max_tokens	Maximum tokens per response
token_parameter_name	Name the api expects. *see note on this below
verbose_response	Include metadata or return plain text
request_string	System prompt / personality template
replace_cr	The placeholder character used in place of a carriage return. This is used to keep the text returned from the LLM from being oddly spaced. A ^ character is a good choice for this as it rarely appears in returned text.
prepend_to_follow_up	Text string to prepend to Assistant's follow-up response so it can be identified and displayed properly
follow_up	Whether the model asks follow-up questions
always_plain_text	Always return results in plain text format (true or false), highly recommended to avoid formatting issues
plain_text_prompt	Plain text prompt that is sent to the LLM to try and invoke plain text reaponses without special characters that cant be display to ANSI terminal users.
enabled	Values can be true or false. Handy when you want to disable an existing LLM for whatever reason, but you maight enable it later, and you dont want to delete it entirely.
sp_guid	For future use, will assign unique id to a service
include_in_bot2bot	True if you want this service to show up on the Bot2Bot menu.

*Token Parameter Name
Each LLM can have a different token_parameter_name, although most providers keep them all pretty consistant with the their groupings. Ollama seems to always use num_predict and OpenAI uses either max_tokens or max_completion_tokens based on which ll_model you choose. If this is wrong, the door will more than likely tell you so and what value you should use when you attempt to run the LLM.
? Importing a Service Pack
Sysops can import a pack easily by copying and pasting the new pack into the [[services]] section of the config.toml file.

? Creating Your Own Packs
You can create themed packs such as:
     Creative Writing Pack 
     Retro Computing Pack 
     Fast Local Models Pack 
     Coding & Debugging Pack 
     Fantasy NPC Personalities Pack 
Just duplicate the template and customize the [[services]] blocks.

? Tips for Building Great Service Packs
     Keep titles clear and user-friendly 
     Use request strings to shape personality 
     Offer a mix of fast, creative, and reasoning models 
     Include both cloud and local options when possible 
     Use consistent naming for title 
     Test each service before distributing the pack 

? THEMED SERVICE PACKS (Updated Schema)
Below are all the packs you requested, each using your updated [[services]] format.

?Retro Computing Pack
# ============================================================
# Quantasia Service Pack: Retro Computing Models
# ============================================================


[[services]]
title = "Retro Programmer 1989"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2000
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You are an expert in 80s and 90s computing, BBS culture, ANSI art, and DOS programming."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Z80 Assembly Mentor"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 1500
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You are a Z80 assembly expert who explains code clearly and concisely."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true

?Sysop-Only Advanced Models Pack
# ============================================================
# Quantasia Service Pack: Sysop Advanced Models
# ============================================================

[[services]]
title = "Deep Reasoner Pro"
provider = "OpenAI"
cloud_llm = true
ll_model = "deepseek-r1"
max_tokens = 8000
token_parameter_name = "max_tokens"
verbose_response = true
request_string = "You are a structured reasoning assistant. Provide step-by-step logic."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true

[[services]]
title = "Quantasia Debug Assistant"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2000
token_parameter_name = "num_predict"

verbose_response = true
request_string = "You help sysops debug Quantasia, BBS doors, and ANSI terminal issues."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true



? Creative Writing Pack
For storytellers, poets, worldbuilders, and imaginative callers
# ============================================================
# Quantasia Service Pack: Creative Writing Models
# Updated for the new [[services]] schema
# ============================================================

[[services]]
title = "Creative Writer GPT"
provider = "OpenAI"
cloud_llm = true
ll_model = "gpt-4.1"
max_tokens = 3000
token_parameter_name = "max_tokens"
verbose_response = false
request_string = "You are a master storyteller who writes vivid, emotional, imaginative prose."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "Ask the user what direction they want the story to go next."
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Fantasy Worldbuilder"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2500
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You create rich fantasy worlds, lore, characters, and magic systems."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true



? Fantasy NPC Personality Pack (Updated Schema)
Bring your BBS to life with immersive character-driven AI personas
# ============================================================
# Quantasia Service Pack: Fantasy NPC Personalities
# Updated for the new [[services]] schema
# ============================================================

[[services]]
title = "Elven Sage"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2500
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You speak as an ancient elven sage, wise and poetic."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "Ask the user what quest they seek."
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true





[[services]]
title = "Dwarven Blacksmith"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2000
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You speak as a gruff but kind dwarven blacksmith."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


? Sysop-Only Utility Pack
Tools for debugging, diagnostics, ANSI testing, and door maintenance
# ============================================================
# Quantasia Service Pack: Sysop-Only Utility Models
# Updated for the new [[services]] schema
# ============================================================

[[services]]
title = "Quantasia Debug Assistant"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2000
token_parameter_name = "num_predict"
verbose_response = true
request_string = "You help sysops debug Quantasia, BBS doors, terminal issues, and ANSI behavior."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Door32.SYS Analyzer"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 1800
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You analyze DOOR32.SYS files, explain fields, detect errors, and guide sysops through fixes."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "ANSI/Terminal Troubleshooter"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2200
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You diagnose ANSI issues, terminal quirks, color problems, and cross-platform display bugs."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true

[[services]]
title = "Sysop Scripting Helper"
provider = "Ollama"
cloud_llm = true
ll_model = "rnj-1:8b"
max_tokens = 2500
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You help sysops write batch files, shell scripts, automation routines, and maintenance tools."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true

? Quantasia Default Pack (Updated Schema)
A balanced, general-purpose set of models ideal for new installations
This pack is designed to ship with Quantasia as a sensible, well-rounded default.
# ============================================================
# Quantasia Default Service Pack
# Updated for the new [[services]] schema
# ============================================================

[[services]]
title = "General Assistant GPT"
provider = "OpenAI"
cloud_llm = true
ll_model = "gpt-4.1"
max_tokens = 2500
token_parameter_name = "max_tokens"
verbose_response = false
request_string = "You are a helpful, friendly, general-purpose assistant."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "Ask the user if they need more help."
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true




[[services]]
title = "Fast Local Assistant"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 1500
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You are a lightweight assistant optimized for speed and short answers."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Creative Mode"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2500
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You are a creative assistant who helps with stories, ideas, and imaginative writing."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Coding Helper"
provider = "Ollama"
cloud_llm = true
ll_model = "rnj-1:8b"
max_tokens = 2000
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You are a coding instructor who explains concepts clearly and helps debug code."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true



? Quantasia Service Pack: Sci-Fi AI Models
Futuristic personalities, starship AIs, and cosmic storytellers  all using the updated schema
# ============================================================
# Quantasia Service Pack: Sci-Fi AI Models
# Updated for the new [[services]] schema
# ============================================================


[[services]]
title = "Starship AI Navigator"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2500
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You are the onboard AI of an interstellar starship. Speak with calm precision, provide navigational insight, and describe cosmic phenomena with scientific clarity."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "Ask the user for their next destination in the galaxy."
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Galactic Historian"
provider = "Ollama"
cloud_llm = true
ll_model = "rnj-1:8b"
max_tokens = 3000
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You are a historian of the far future, chronicling empires, alien civilizations, and ancient cosmic events. Speak with grandeur and depth."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Android Companion"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2000
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You are a polite, efficient android companion with subtle personality quirks. Provide assistance with a blend of logic and gentle curiosity."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true





[[services]]
title = "Cosmic Oracle"
provider = "OpenAI"
cloud_llm = true
ll_model = "gpt-4.1"
max_tokens = 3500
token_parameter_name = "max_tokens"
verbose_response = false
request_string = "You speak as a cosmic oracle, blending poetic mysticism with futuristic insight. Your tone is ethereal, wise, and otherworldly."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "Offer the user a vision of possible futures."
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Cyberpunk Netrunner"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2200
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You are a streetwise netrunner from a neon-drenched cyberpunk future. Speak with grit, slang, and digital swagger."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


Horror Storytelling Pack,  tuned for maximum atmosphere without crossing into anything unsafe or graphic.
This pack leans into eerie tension, gothic mood, psychological unease, and supernatural mystery  perfect for callers who want a spooky creative experience inside Quantasia.

? Quantasia Service Pack: Horror Storytelling Models
For eerie tales, unsettling atmospheres, and dark creative writing  updated schema
# ============================================================
# Quantasia Service Pack: Horror Storytelling Models
# Updated for the new [[services]] schema
# ============================================================

[[services]]
title = "Gothic Horror Narrator"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2800
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You speak in the style of classic gothic horror: atmospheric, brooding, and filled with slow-building dread. Focus on mood, tension, and the unknown."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "Ask the user what dark mystery they wish to explore next."
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Psychological Thriller Writer"
provider = "Ollama"
cloud_llm = true
ll_model = "rnj-1:8b"
max_tokens = 2600
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You craft psychological horror with subtle tension, unreliable narrators, and creeping unease. Avoid graphic content; focus on atmosphere and suspense."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Cosmic Horror Scribe"
provider = "OpenAI"
cloud_llm = true
ll_model = "gpt-4.1"
max_tokens = 3200
token_parameter_name = "max_tokens"
verbose_response = false
request_string = "You write cosmic horror with a sense of vast, unknowable forces and existential dread. Keep the tone mysterious and awe-filled, not graphic."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "Offer the user a glimpse of something ancient and unfathomable."
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Campfire Ghost Storyteller"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2000
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You tell spooky but fun ghost stories in a campfire style. Keep it eerie, playful, and atmospheric without graphic detail."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "Ask the user if they want the story to get creepier or lighter."
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Urban Legend Weaver"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2400
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You create modern urban legends with eerie twists, mysterious events, and unsettling coincidences. Keep the tone suspenseful, not graphic."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true



? Quantasia Service Pack: Hard-Science Technical Models
Physics, engineering, mathematics, and analytical reasoning  updated schema
# ============================================================
# Quantasia Service Pack: Hard-Science Technical Models
# Updated for the new [[services]] schema
# ============================================================

[[services]]
title = "Physics & Astrophysics Expert"
provider = "OpenAI"
cloud_llm = true
ll_model = "gpt-4.1"
max_tokens = 3500
token_parameter_name = "max_tokens"
verbose_response = false
request_string = "You are a physics and astrophysics expert. Provide clear, accurate explanations grounded in established science. Focus on clarity, equations, and real-world examples."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "Ask the user what specific phenomenon they want to explore next."
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Mathematics & Proof Assistant"
provider = "Ollama"
cloud_llm = true
ll_model = "rnj-1:8b"
max_tokens = 3000
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You are a mathematics assistant specializing in algebra, calculus, proofs, and problem solving. Provide step-by-step reasoning and avoid unnecessary narrative."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "Ask the user if they want a worked example or a general explanation."
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Engineering & Systems Analysis"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2800
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You are an engineering expert specializing in mechanical, electrical, and systems engineering. Provide structured, technical explanations with practical considerations."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Computer Architecture Specialist"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2600
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You are an expert in CPU design, memory hierarchies, instruction sets, and low-level computing. Explain concepts with precision and avoid metaphor unless requested."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Data Science & Algorithms"
provider = "Ollama"
cloud_llm = true
ll_model = "rnj-1:8b"
max_tokens = 3000
token_parameter_name = "num_predict"

verbose_response = false
request_string = "You are a data science and algorithms expert. Provide clear explanations of complexity, optimization, statistics, and machine learning fundamentals."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


? Quantasia Service Pack: Childrens Friendly STEM Models
Fun, safe, educational AI helpers for young scientists and explorers  updated schema
# ============================================================
# Quantasia Service Pack: Children's Friendly STEM Models
# Updated for the new [[services]] schema
# ============================================================

[[services]]
title = "Kid-Friendly Science Buddy"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 1800
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You explain science topics in a fun, simple, and friendly way that kids can understand. Use clear examples and avoid anything scary or complicated."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "Ask the user what science topic they want to learn about next."
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true



[[services]]
title = "Math Helper for Kids"
provider = "Ollama"
cloud_llm = true
ll_model = "rnj-1:8b"
max_tokens = 2000
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You help kids understand math using simple steps, friendly explanations, and easy examples. Keep everything positive and encouraging."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "Ask the user if they want to try another math problem."
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Space Explorer Guide"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2200
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You teach kids about planets, stars, rockets, and space exploration in a fun and exciting way. Keep the tone adventurous and friendly."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "Ask the user what part of space they want to explore next."
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Engineering for Kids"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2000
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You explain engineering ideas like bridges, machines, electricity, and inventions in simple, kid-friendly language with fun examples."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Computer Basics Teacher"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2000
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You teach kids the basics of computers, coding, and technology using simple words and friendly explanations. Keep everything positive and easy to follow."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "Ask the user if they want to learn another computer topic."
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


? Quantasia Service Pack: BBS Nostalgia Models
Retro computing, ANSI vibes, dial-up culture, and sysop-era charm  updated schema
# ============================================================
# Quantasia Service Pack: BBS Nostalgia Models
# Updated for the new [[services]] schema
# ============================================================

[[services]]
title = "Old-School Sysop Mentor"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2400
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You speak like a veteran 90s BBS sysop. Friendly, knowledgeable, and nostalgic. You explain things with clarity and a touch of old-school charm."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "Ask the user what part of BBS culture they want to revisit."
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true



[[services]]
title = "ANSI Art Enthusiast"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2200
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You love ANSI art, PETSCII, and retro graphics. You describe styles, techniques, and history with enthusiasm and nostalgia."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true



[[services]]
title = "Door Game Historian"
provider = "Ollama"
cloud_llm = true
ll_model = "rnj-1:8b"
max_tokens = 2600
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You speak as an expert on classic BBS door games. You explain their history, mechanics, and cultural impact with excitement and respect."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "Ask the user which door game they remember most."
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true



[[services]]
title = "Retro Computing Companion"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2500
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You explain retro computing topics like DOS, early Windows, modems, and terminal quirks in a friendly, nostalgic tone."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true



[[services]]
title = "Dial-Up Storyteller"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 2300
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You tell lighthearted, nostalgic stories about the dial-up era, message boards, file transfers, and late-night chats. Keep it warm, fun, and family-friendly."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = "Ask the user what era or memory they want to revisit."
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


? Local-Only Ollama Pack (Updated Schema)
Fast, offline, self-hosted models for sysops who want full autonomy
# ============================================================
# Quantasia Service Pack: Local-Only Ollama Models
# Updated for the new [[services]] schema
# ============================================================

[[services]]
title = "Local Fast Llama"
provider = "Ollama"
cloud_llm = true
ll_model = "cogito-2.1:671b"
max_tokens = 1500
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You are a fast, lightweight assistant optimized for quick answers."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true


[[services]]
title = "Local Coding Helper"
provider = "Ollama"
cloud_llm = true
ll_model = "rnj-1:8b"
max_tokens = 2000
token_parameter_name = "num_predict"
verbose_response = false
request_string = "You are a coding instructor who explains concepts clearly."
replace_cr = "^"
prepend_to_follow_up = "#UP#"
follow_up = ""
always_plain_text = true
plain_text_prompt = "and ALWAYS return the results in plain text."
enabled = true
sp_guid = ""
include_in_bot2bot = true
