STOBE

Dwemer Dynamics Home Overview Installation Configuration Roleplay Settings Kenshi Settings Large Language Models Text-to-Speech Modders Guide FAQ Remote Hosting Guide Linux Setup Guide Log Files

⬇️Download 🎥Video Guide DiscordDiscord YouTubeYouTube PatreonPatreon

STOBE Configuration

This page explains the main STOBE setup pages inside StobeServer in plain language. Use it after installation when you want to tune how STOBE behaves across your playthrough.

Global Settings

What this page does: This is the main global behavior page for StobeServer. Changes here affect the overall AI behavior for the whole save, not just one NPC.

  • Prompting: This is where StobeServer stores the main prompt head, emote mood list, general instructions, and roleplay instructions.
  • Rechat: Controls how STOBE chooses follow-up speakers, including rechat mode, strict targeting, and whether the original player speaker can be selected again.
  • Memory: Covers the global memory system and how much older context STOBE keeps bringing back into play.
  • World Knowledge: Includes broader world knowledge behavior, including whether race knowledge should be injected automatically when matching entries exist.
  • Relationship System: This is where the shared relationship layer is turned on or off for server-side interaction tracking.
  • Player Faction: Lets you define a custom player-faction name and an optional extra prompt block for how that faction should be treated.
  • Context Selections: Lets you choose which context blocks and subsections STOBE includes in system prompts.

Profiles

What this page does: Profiles are reusable behavior presets for NPCs and factions. Instead of hand-tuning everyone one by one, you build profile setups and assign NPCs to them.

  • Basic profile setup: Each profile has a label, can be marked as a default NPC profile, and can also be marked as a player-faction profile.
  • Main prompt text: Profiles carry both a Prompt Head and a Profile Prompt, which shape the behavior style of NPCs using that profile.
  • Connector choices: Each profile can point to its own Response, Diary, Autochat, Memory, Background Life, Dynamic Profile, Relationship, and TTS connectors.
  • Dynamic profile behavior: Profiles decide whether STOBE can rewrite fields like personality, speech style, and goals over time.
  • Diary behavior: Profiles also control auto-diary behavior, including how often entries can be written and how many events are needed first.
  • Workflow tools: The page supports cloning, exporting, importing, profile testing, and profile-assignment rules for imported NPCs.

API Keys

What this page does: This is the shared API key page for StobeServer. You save provider keys here once, then the connector pages reuse them.

  • Preset providers: The current built-in presets include OpenRouter, OpenAI, Groq, Nano-GPT, Google, Cartesia, and Inworld.
  • Used across the server: These saved keys are then selected by the Large Language Model and Text-to-Speech connector pages.
  • Custom key rows: You can also add your own named API badge entries for other supported services.
  • Practical use: If a connector needs a cloud login, this page is usually where that key gets stored first.

Narrator

What this page does: This page controls the STOBE narrator layer, including when it speaks, what profile it uses, and how it presents itself.

  • Core toggles: You can turn the narrator on or off, enable welcome narration, and enable random narration.
  • Chance and cooldowns: The page controls welcome cooldown, random narration chance, and random narration cooldown.
  • Profile and voice: You choose which profile the narrator uses, its voice ID, and which world-knowledge tags it should pull from.
  • Prompt override: There is also a narrator-specific prompt-head override if you want the narrator to follow special instructions.
  • Dynamic profile updates: The narrator can also use dynamic updates for personality, speech style, and goals if you turn that on.
  • Narrator identity: Core summary, background, personality, speech style, and goals are all editable on this page.

Action Editor

What this page does: Action Editor controls which server-side actions STOBE is allowed to use and whether they are active.

  • Enabled state: You can turn individual actions on or off without changing the shipped base rows.
  • Search and filtering: The page supports live search plus enabled and disabled filters so you can find specific actions faster.
  • Base versus custom: STOBE keeps track of which actions are still stock and which ones now have custom override rows.
  • What you are editing: The list shows the command name, visible action name, description, and current enabled state.
  • Practical use: This is mainly the page to visit when you want to stop a specific AI action from being used or re-enable one that is off.

Prompts Manager

What this page does: Prompts Manager is the page for editing STOBE's built-in prompt text one prompt at a time.

  • Default versus custom: Each prompt key keeps its shipped default text and an optional custom override.
  • Edit and clear: You can write a replacement prompt for any key, or clear it to fall straight back to the default.
  • Description: The page also shows a short description for what each prompt is for before you change it.
  • Import and export: Custom prompt changes can be exported as CSV and imported again later for backup or sharing.