September 7 Update

September 7 Update

Hi everyone! This Steam update brings a redesigned companion page, Ruka's Lofi study room, and local and online video imports. We've also replaced the old usage-count display with points, refreshed settings and onboarding, and continued improving explanations and multiplayer.

A new companion page

• Companions now have a dedicated showcase with a starry backdrop, rising particles, and floor reflections. Choose a background color theme and save it locally for each character.
• Search the roster, sort by name or recent updates, and switch characters with the mouse wheel, horizontal dragging, or arrow keys. Card colors distinguish official, Workshop, and locally created characters.
• Rotate 3D characters and use the wheel to move the camera closer. Skins, expressions/actions, and voice tabs have their own close-up views, while illustrated characters gain a perspective-tilt effect. Model preloading, clipping, and reflections have also been adjusted.
• Browse profiles, skins, expressions/actions, voice previews, and CGs in separate tabs. Click a CG for a full-screen preview. Character creation and editing now use this page instead of the old form.
• Workshop controls bring together item pages, publishing, and updates. Borrowed characters can be cloned into local copies, and published characters cannot be updated again without new changes. Permanently deleting a local character requires an additional confirmation.

Ruka's Lofi study room

• Select the new study-room mode, then click Start to enter. Its background animation and music are stored locally, so these assets do not need to load online.
• A collapsible playlist gives all 50 tracks names and lets you select, play, or pause a track directly.
• A Pomodoro timer lets you customize focus and break durations and remembers them locally. Sound and visual reminders mark the end of a session; duration settings open in a small separate dialog.

Import documents or videos

• The start screen is now a dedicated full-screen page with preview images for Story, Free Mode, and Lofi. Choose either a document or video input, with extra options alongside it and a steadier layout when switching inputs.
• Import local videos or online sources such as Bilibili and MP4 links. Speech is transcribed into text for subsequent learning. Local imports accept MP4, MOV, WebM, and MKV files up to 500 MB and 60 minutes long.
• The Blackboard Mode option is now Image Parsing Mode, which analyzes document figures in advance. Story blackboards can still appear for formulas and key points when useful, without enabling image parsing.
• Image explanations are no longer limited to one representative figure: generation aims to use more relevant images where they fit the discussion, without forcing images in just to increase coverage. Instructions also place more emphasis on matching explanations to the figure's labels.

Free Mode and story explanations

• Paper Q&A is now called Free Mode. After a document is read, you can keep asking questions and toggle spoken replies as needed.
• Character dialogue has been adjusted toward everyday conversation, with fewer report-style openings, repeated summaries, and forced affectionate phrases. Simple questions can receive short answers.
• Blackboard text is no longer appended to Free Mode dialogue. Source labels no longer interrupt the answer, while source buttons still let you check the original text. The outdated usage-count note beneath the input has been removed.
• Improved character-name matching, formula display, and blackboard use for important definitions, derivations, and key points in stories.

Points and today's spending

• Built-in AI usage is now displayed as points. Check your remaining balance through the profile card in the top-right corner.
• New daily spending details and pricing rules show which feature used your points. Story generation and voice are charged separately; video and image parsing are additional items. Refer to the displayed daily allowance and prices for current values.
• Turning on voice in Free Mode does not itself cost points. Charges apply when a voiced reply is generated.

Easier Steam multiplayer

• Join Room now searches for joinable Steam friends' rooms inside the game. Click a result to join instead of simply opening the Steam friends window.
• Character artwork and other shared assets can resume interrupted transfers, retry failures, and recover from stalls, helping downloads continue through unstable connections.
• Ruka's multiplayer reactions have been refreshed with 12 selected transparent-background stickers.

Menus, settings, and onboarding

• Settings now use a warm off-white, full-screen layout with reorganized categories, stronger text contrast, and refreshed sliders and switches.
• A new Display Settings category groups visual options and includes overall UI scaling from 80% to 120%.
• Updated onboarding covers the new mode selection, imports, points, and settings. It uses a built-in example without spending points. Opening splash screens can also be skipped one at a time with a click.
• Improved the Russian dialogue font, filled in more interface translations, and fixed the profile avatar and some menu buttons being obscured by characters.

Loading, recommendations, and reliability

• Startup reads locally saved custom and Workshop characters first, while subscription refreshes run in the background, reducing cases where characters remain unavailable while Workshop checks finish.
• Daily recommendations are saved locally, so the same preferences do not need repeated requests that day. New selections also try to avoid recently recommended documents.
• Improved custom AI endpoint compatibility and some generation-error messages. Reference-audio uploads check the 20-second limit so overly long clips can be caught before voice generation.
• Added local runtime logs with automatic cleanup of older records and masking of common sensitive fields to help diagnose problems.

Thanks for all your feedback! If you run into display, explanation, or multiplayer issues, please share a screenshot and steps to reproduce them through the Steam Community, Discord, or our QQ groups.