NextNeural/ARIA
Mapa prywatnego backendu nie nalezy do publicznego Space'a
aria(mobile): long-press menu is plain text (no box) left-aligned to bubble — time above, Odpowiedz below; css/js v23
aria(mobile): show each message's sent date+time (UTC) above 'Odpowiedz' on long-press; css/js v22
aria(mobile): react-hint is now a line-art smiley+ icon (stroke, like the paperclip), not a colored emoji; css/js v21
aria: block = silence + "can't reply" composer + cheeky unblock comeback; mute unchanged; raise limit to 50; drop usage countdown
aria: profile popup (tap header avatar → swing-in modal, animated close); peek panel routing fix note. css/js v19
aria(mobile): faint 🙂+ react-hint on the latest ARIA message (visual only, vanishes once reacted); remove the tip line; css/js v18
aria(db): readable chronological transcript view (email · time · sender · text); extract user text from JSON, UTC time, skip system_idle
aria: profile-picture avatar; come-back-after-limit catch-up; email in usage + read-by-email views
aria: examples-are-examples (less robotic repeats); fix missing proactive pings; unify reply timing; drop ↩ from reply menu
aria(mobile): hide the redundant tip + tighten input while keyboard is open; css v16
aria(mobile): keep keyboard open on Send; collapse top bar while typing; drop the input-area divider line (tighter, more messages visible); css v15 / js v15
aria: interruptible delivery — adapt when the user cuts in mid-reply
aria(mobile): JS fallback to kill iOS double-tap zoom (non-passive touchend preventDefault); js v14
aria(mobile): reply bar uses a clean white line arrow (SVG) instead of the ↩ emoji; css v14
aria(mobile): touch-action:manipulation on rows/bubbles — stop double-tap-to-zoom on double-tap-❤️ (iOS+Android); css v13
aria: fix iOS WS dropout + reply-to stripping; double-tap heart; drop swipe arrow
aria: iOS keyboard viewport fix + more reactions/reply-to/bubbles + time awareness
system prompt update
aria: reconcile reaction set (❤️😆😮🥹😢😡) across worker+brain+client; instant keyboard fit; brb-suppression + model-judged idle; emotional-tone continuity; tip under input bar
aria(mobile/ui): fix keyboard layout jump (lock page + recompute through anim); swipe-right-to-reply; truncate quote >25 chars to 20+...
TEST persona update v2
TEST persona.py update
Update cloudflare/pages-integration/js/aria-chat.js
inne emotki
aria: smart clustering (wait for multi-bubble buildup), per-bubble relevance/threading, reply-to flow fixed, interchangeable XD/hahah + hmm/wtf/no wayyyy
aria: fix instruction-leak (frame+lang moved into system msg, convo ends on real user turn); react ~every 3rd msg; 'byt'->'inteligencja'; 400-char limit
aria: reply delay per bubble -> 1-8s; <5 words = 1s flat
aria(fix): no-response bug — relax over-aggressive echo filter (was dropping all bubbles when aya reused a phrase); retry + fallback so a turn is never silent
aria: reply delay per bubble 7-19s -> 2-9s (still length-scaled)
aria huge update: reply, hold-to-react (blur+picker+menu+emoji-fly), live sync, 100-char, spam guard, limit sequence, 7-19s/bubble, 15-50s idle, windowed memory
Update brain/orchestrator.py
brain: switch to aya-expanse-32b (your 32B/gemma-4 weren't available; gemma-4 broke Aria entirely)
aria(mobile): full-bleed chat — override shared .app-body centering/padding (removes side bars); css v9
zmiana na google/gemma-4-31b-it
fix zmiany modelu
Model change to Qwen/Qwen2.5-14B-Instruct
Update sys prompt
update prompt
Tescik 2
Update prompt - role.
Tescik
Update system promptu
Nowy prompt
aria: recognize user gender from their name + address by name sometimes
aria: warmer/more caring persona (not overdone); mobile chat full-bleed on all phones
aria: stop leaking internals — event ids + mechanics no longer reach the user
brain/persona: restore multi-bubble splitting (lost in the 'smarter' rewrite)
aria(ui): livelier typing dots (bigger bounce + scale, brighter), cache-bust css
brain/persona: playful BUT intelligent — stop playing dumb
