Settings & Preferences
Sholarel's Settings page collects every personalization option, so you can tune the app to your own habits and needs. Open Settings to view and change them all.
Appearance
Appearance covers interface themes, fonts, character models and avatars. See these dedicated pages:
- Themes & Appearance — themes, fonts, ambient atmosphere
- Character Model — VRM/Live2D models, avatars, floating window
Preferences
Interface Language
Sholarel supports two interface languages, Simplified Chinese and English.
- Go to Settings → Preferences → Language
- Pick a language from the dropdown
- The interface switches immediately, no restart needed
Main Chat Model Genesis
Genesis users can switch the AI model used for main chat:
| Model | Strengths |
|---|---|
| Gemini 3.1 Flash Lite | Light and fast, great for casual chat |
| Gemini 3.5 Flash | Balanced performance and quality (default) |
| Claude Opus 4.7 (1M) | Strongest reasoning, very long context |
| Claude Haiku 4.5 | Fast responses at a lower cost |
Choosing your own main chat model is a Genesis-only privilege. Other tiers use the system default model.
Gemini Thinking Depth
Controls how deeply Gemini reasons before replying:
| Level | Effect |
|---|---|
| minimal | Fastest responses, suited to simple conversation |
| low | Light reasoning, balancing speed and quality |
| medium | Default level, suits most situations |
| high | Deep reasoning, first choice for complex questions |
Higher levels reason better but respond more slowly and use more quota. Applies to Gemini models only.
Sub-Agent (AI Agent) Genesis
When enabled, the AI sub-agent can carry out complex multi-step tasks such as running terminal commands, browsing the web and automating your desktop. See AI Agent.
Gender
Set your gender to help AI understand your identity and communication preferences better. You can change it at any time.
Reply Pacing
On by default. AI renders replies sentence by sentence, imitating a natural typing rhythm. Turn it off to shorten the gaps between sentences and show messages faster.
Proactive Reply Delay
When AI has something to say on its own (a reminder, a check-in), this sets the range of delay:
- Minimum delay: how many seconds at the earliest (10 seconds by default)
- Maximum delay: how many seconds at the latest (60 seconds by default)
Tuning this range makes AI's proactive messages feel more natural — neither an instant reply nor a long wait.
Launch at Startup (Desktop)
- Launch at startup: start Sholarel automatically when the system boots
- Start minimized: run minimized on startup so it doesn't interrupt boot
Desktop Widgets (Desktop)
When enabled, floating windows on your desktop show summary, to-do, habit and mood widgets at all times. See Life Feed & Widgets.
Pseudo-Fullscreen While Focusing (Desktop)
Entering Focus Mode expands the window into pseudo-fullscreen to reduce distractions. See Focus Mode.
Half-Body Video Calls (Desktop)
During video calls, the AI character switches to half-body mode with default motions. Expressions, lip-sync and cursor following still work. On by default.
Auto-Hang Up After 60 Minutes
On by default. At 60 minutes into a video call a reminder appears, and the call hangs up automatically 5 minutes later, so a forgotten call doesn't waste quota. See Video Call.
Sensitive-Area Rendering Protection
On by default. When the camera angle moves close to a sensitive area, the AI character automatically applies blur and shake protection.
Emoji Reference Image
AI needs a reference image to determine the style when generating emoji packs:
- Default: captured automatically from the current VRM model
- Custom: turn on "Use a custom reference image" and upload your own
See Emoji & Stickers.
Floating Window Size (Desktop)
Adjusts the size of the desktop floating character, from 50% to 200%, 100% by default.
Live2D Model Scale
Adjusts how large the Live2D model appears, from 50% to 200%, 100% by default.
Subscription
Check your current tier, remaining days and Echo quota. See Subscription & Benefits.
TTS Voice Settings
Adjust the emotional intensity and reference audio for AI's voice. See TTS Voice Settings.
QQ Bot
Configure credentials for a QQ Open Platform bot. See QQ Bot.
WeChat Integration
Connect your WeChat account so AI can reply to WeChat messages for you. See WeChat Integration.
Alarms & Reminders
Manage alarms and scheduled reminders. See Alarms & Reminders.
MCP Tool Extensions
Configure MCP servers to extend AI's capabilities. See MCP Tool Extensions.
Danger Zone
At the bottom of Settings, the Danger Zone holds irreversible actions:
| Action | Description |
|---|---|
| Clear local cache | Clears cached messages and data on this device; cloud data is untouched |
| Reset all settings | Restores every setting to its default |
Actions in the Danger Zone cannot be undone — proceed carefully.
Tips
- Most settings take effect immediately, no restart required
- Settings sync to the cloud automatically, so they stay consistent across devices
- If something behaves oddly, use "Reset all settings" in the Danger Zone