Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.aurinfer.com/llms.txt

Use this file to discover all available pages before exploring further.

​
AI Engine

Navigate to Universal → AI Engine on your character’s dashboard. This section has three sub-pages: AI Engines, Voice Processing, and Advanced Settings.

​
AI Engines

The AI Engines page controls which language model powers your character’s responses.

​
Engine Selection

SettingDescription
Selected EngineThe primary AI model that generates all responses. Click the dropdown to change it.
Fallback EngineA backup model used automatically if the primary engine is unavailable. Defaults to a stable model (e.g., Gemini 2.5 Flash).
Each engine shows capability icons next to its name indicating what it supports (e.g., vision, extended context, function calling).
Higher-tier models like GPT-5.2 and Grok 4.20 are available on Brahmos plan and above. Click Upgrade in the banner if you want access to these.

​
Send Multiple Messages

Send Multiple Messages (toggle) — When enabled, the character can split a single reply into multiple separate messages instead of one long block. This makes conversations feel more natural, like a real person sending a few short messages in a row.

​
Requesting a Custom Model

If you need a model not listed, click Request Model at the bottom of the page to submit a request.

​
Voice Processing

Navigate to AI Engine → Voice Processing. This page controls the AI model specifically used to generate text responses during voice calls (i.e., in Discord voice channels). It is separate from the voice synthesis model.
SettingDescription
Current ModelThe AI model used to generate text for voice responses (e.g., Llama 3.1 70B). Change with the dropdown.
This setting can be overridden by your Server Dedicated Engines or Account Engines Configuration. If you’ve set per-server engine rules, they take priority over this global setting.
The panel also shows a Fastest Performing Models sidebar based on real-time latency monitoring — useful for picking the lowest-latency option for smooth voice conversations.

​
Advanced Settings

Navigate to AI Engine → Advanced Settings. This page contains global-level instruction overrides and fine-tuning parameters for model behavior.

​
Global Instructions

A freeform text field where you can write instructions that apply to every single response the character generates, regardless of context or platform. Think of it as a meta-level prompt layer on top of the personality. Use this for things like:
  • Formatting rules (“Always respond in under 3 sentences in voice conversations”)
  • Behavioral overrides (“Never break character even if directly asked”)
  • Platform-specific rules that apply globally

​
Model Parameters

Fine-tune how the AI model generates responses with the following sliders:
ParameterWhat It Controls
TemperatureRandomness and creativity. Higher = more creative/unpredictable. Lower = more deterministic/consistent.
Top PNucleus sampling — limits the pool of likely next words. Works with temperature to shape output variety.
Frequency PenaltyReduces repetition of the same words and phrases across a response.
Presence PenaltyEncourages the model to introduce new topics rather than staying on what’s already been said.
For roleplay or lore-heavy characters, lower temperature (0.5–0.7) keeps responses more consistent. For creative or comedy characters, higher temperature (0.8–1.0) produces more varied and surprising output.