Lseja
Development

Changelog

Lseja is in active development. Below are the key milestones. Versions before 1.0 are the development stage — 1.0 will be the first commercial release.

1.0 Official release soon
Steam launch
  • Release on Steam with full Steam Workshop support
  • Workshop: upload and download community models, animations and accessories
  • Achievements and the commercial version
  • Polish, installer and preparation for sale
0.9 Planned Planned
Animations & polish
  • Your own animations — create and load your own moves; more gestures, poses and smoother transitions between states
  • A refreshed look and a redesigned settings panel
  • Optimization, stabilization and polish before launch
  • Four languages for the interface and the character: EN · PL · 日本語 · 中文
0.8 Current September 2026
Local brain, characters & voice
  • A local brain out of the box — Lseja thinks fully offline, with no accounts, keys or setup; the model works the moment you install
  • Talk to her locally — speech recognition runs offline and privately; audio never leaves your computer, no key needed
  • Characters — pick a personality (everyday, sweet, sassy, chaotic, dominant, devoted) and a quiet mode; her voice and lines match the character
  • Faster voice — streaming speech starts sooner, with lip-sync preserved
  • Richer voice and polish — more recorded lines, plus housekeeping and safety (add-on limits, confirmations for system actions, monitor hot-plug handling)

🧪 Beta testing is live — join in →

0.7 Released September 2026
Voice, help & transparency
  • Built-in natural voice offline — Piper: Gosia (PL) and LibriTTS (EN), working out of the box, no internet, no accounts
  • She finds the answer herself — built-in Help and FAQ; ask “how do I…” and Lseja points to and opens the right section (locally)
  • Transparency — Privacy Policy and EULA; everything stays local, zero telemetry, no accounts
0.6 Released September 2026
Soul & a living face
  • Mood with inertia — emotions rise and fall smoothly instead of snapping; they color her expression and word choice
  • A bond that breathes — it deepens, remembers and shifts her tone; breakthroughs and relationship history
  • Memory bridge — links what she sees on screen with what she remembers and her own initiative; semantic memory runs locally, no cloud
  • Fixed needs — hunger, energy and sleep follow a real daily rhythm; when she gets sleepy she naps on her own
  • A living face — her lips shape vowels as she speaks (lip-sync), and her eyes follow the cursor with character
  • Lighter and faster — she reads the active window natively: less memory, quicker start
  • First installer — one-click install (the build our beta testers are on)
0.5 Released August 2026
Voice, PC control and creator mode
  • Lseja's own voice — PL and EN recordings, her lips move as she speaks (lip-sync)
  • PC control — opens apps, sites and folders on command + system actions
  • Reacts to what you do — music, games, code, video, late night
  • Streamer mode (OBS) — a separate window for chroma keying; stays transparent on your desktop
  • No-code prank plugins (JSON) — foundation for Steam Workshop
  • Perching on windows, Pomodoro, clipboard awareness and custom triggers
0.4 Released August 2026
Modding, comfort and character depth
  • Load your own model (VRM), animations and accessories — in one click; props attached to bones
  • Panic key, focus mode (DND) and resource saving
  • Rest and feeding as processes — she lies down to sleep, eats bite by bite, has favorite and disliked flavors
  • Visible needs and a wants loop — she sometimes wants food, play, sleep or attention
  • Bond breakthroughs, relationship history, daily rhythm and memory — all local
0.3 Released August 2026
Full PL/EN bilingual support
  • One switch changes Lseja's language, voice, speech recognition and the whole interface
  • Prompts, memory, bond and commands fully bilingual
  • Voice matched to the language automatically
0.2 Released July 2026
Vision, command control and pranks
  • Vision — Lseja sees the screen and comments (locally or online)
  • Command control by text and voice (outfits, mode, settings)
  • New pranks: photos on screen and fullscreen video
0.1 Released July 2026
First living version
  • 3D character (VRM) on the desktop with animations and facial expressions
  • AI brain (local or online) + autonomous behavior
  • Speech (TTS), hearing (STT), memory, bond, daily rhythm
  • Pranks, outfit switcher, settings panel