Skip to content

Settings Reference

AMousePad edited this page Apr 14, 2026 · 1 revision

Settings Reference

1776153517724

This page covers everything in the Advanced page.

All settings are per-user.

Refinement section

Setting Default Range Purpose
Show Diff After Refine On - Show a word-level diff modal after every successful refinement. Turn off for silent background
refinement (auto-refine pairs well with this off).
Min Char Threshold 20 0–500 Skip refinement for messages shorter than this. Saves tokens on trivially smallmessages.
Bulk Delay (ms) 2000 0–30000 Milliseconds to wait between messages during "Hone All" runs. Keeps you from hitting rate
limits on a 50-message chat.
Request Timeout (seconds) 120 30–600 Per-LLM-call timeout. Applied to every stage. A 3-stage pipeline can take up to 3× this
before the whole refine aborts.
Play Sound on Refinement Complete Off - Audio cue when a refine finishes.
Custom Sound URL "" - Paste any URL to override the default sound. Visible only when Play Sound is on.

Widget section

Setting Default Range Purpose
Hide Widget Off - Hide the floating chibi/icon widget entirely. Drawer plus per-message button still work.
Confirm Widget Taps Off - Require a second tap within 4 seconds before the widget fires. Useful on mobile to avoid accidental triggers.
Widget Size (px) 48 24–1920 Widget diameter. Also exposed via the widget's long-press / right-click context menu.
Lumia Mode On - Render the chibi character sprites. Off uses a classic circular icon pill.

Debug Logging section

See Multi-User and Privacy for the privacy story.

Setting Default Range Purpose
Enable Debug Logging Off - Start capturing per-user debug entries to an in-memory ring buffer. Off by default.
Max Log Entries 2000 100 - 20000 Ring buffer capacity. Resizing preserves the most recent entries.
Record Full LLM Payloads Off - Include text sent to and received by the LLM in debug logs when honing.

Plus two action buttons:

  • Copy Debug Logs. Formats the buffer with timestamps and copies to clipboard. Paste into a bug report.
  • Clear Logs. Empties the buffer without disabling capture.

A "Buffer Status" row shows current count, capacity, and whether logging is on.

Drawer-tab settings (not on this page)

For completeness, settings that live elsewhere in the UI:

  • Active output preset and active input preset. Drawer -> Output / Input preset bar.
  • Active model profile. Drawer -> Models tab.
  • Auto-Refine AI toggle. Drawer -> Hone tab -> Automation section.
  • AI Message POV / User Message POV. Drawer -> Output / Input.
  • Max Lorebook Tokens / Max Message History Tokens. Drawer -> Output (or Input) -> Context sub-subtab.

Next

Clone this wiki locally