Conversations

Side panel & chat settings

The six-tab panel that lives beside every chat — character info, gallery, scenarios, memory, plugins, and the per-chat Advanced Settings where you tune the model, voice, display, and data.

Everything that isn't the message stream lives in the side panel — a six-tab strip you open from the chat. On desktop it slides in beside the conversation; on mobile it opens as a sheet. The conversation keeps running while it's open.

┌─────────────────────────────────────┐
│  ℹ️  🖼️  📖  🧠  🧩  ⚙️           │  ← six tabs
├─────────────────────────────────────┤
│                                     │
│   (selected tab's content)          │
│                                     │
└─────────────────────────────────────┘

The six tabs

Tab	Icon	What's inside	Deep dive
Info	ℹ️	The character's public profile — bio, tags, creator, sample of how they talk.	Finding characters
Gallery	🖼️	The character's images: avatar, reference art, and any gallery shots the creator published.	Image guidelines
Scenarios	📖	Pre-built contexts the creator authored (e.g. "Coworker", "Rival"). Switch the framing without leaving the chat.	Scenarios
Memory	🧠	View, edit, and pin what the character remembers about you.	Memory
Plugins	🧩	Plugins active in this chat — prompt injectors, tools, and interactive buttons.	Plugins
Settings	⚙️	Per-chat Advanced Settings — model, voice, display, data. Covered below.	This page

The first three tabs are about the character; the last three are about your conversation with them.

Advanced Settings (⚙️)

The Settings tab is where you tune how this specific chat behaves. Changes apply to the current conversation, not globally. It's grouped into four sections.

These settings are per character / per chat. The model you pick for one character doesn't change what another character uses.

Model & Generation

Setting	What it does
AI Model	Opens the model picker. Each model shows quality (real user A/B win-rate), speed (recent latency), and price multiplier. Choosing a model →
Reasoning Mode	Let the model think step-by-step before replying — higher quality, slower, more tokens. Some models are always-on. Reasoning mode →
NSFW Content Intensity	Three levels — Literary, Moderate, Explicit. Hidden when the creator has defined a custom NSFW prompt for the character.
Narration Style	First person ("I"), Third person ("he/she/they"), or AI Decide.

Voice

Setting	What it does
Voice	Pick the TTS voice used when messages are read aloud. Free voices use Edge TTS; Premium voices use MiniMax. Tap to preview before applying.
Auto-play Voice	When voice mode is on, automatically play TTS for new replies. Turn it off to keep voice-style replies without audio — you can still tap any message to play it. Voice call mode always plays audio regardless.

Full voice guide →

Display & Interface

Setting	What it does
Language	The reply language for this chat. The interface follows your selection.
Reply Suggestions	Toggle the smart-suggestion chips above the input. Smart suggestions →
First Response Enhancement	Show or hide the six-button panel under the first AI reply. First response →
Alternative Response	Allow the occasional side-by-side dual-model first reply. Dual comparison →
Reasoning Pane	Show the AI's thinking pane, and whether it auto-expands.

Data Management

Action	What it does
Chat History	Browse every past conversation you've had with this character.
Start New Session	Begin a fresh thread while carrying over the context and memories from the current one — useful when a chat gets very long.
Import Chat	Import a conversation from a JSONL file (SillyTavern-compatible).
Export Chat	Export this conversation as JSONL (SillyTavern-compatible) or plain text, for backup or sharing.

Chat overview — the whole chat screen mapped
Memory — how the memory tab works in depth
Plugins — what lives in the plugins tab