fix: analytics only tracked after successful GPU generation + indent fix

feat: returning user IP recognition on load - skip ToS for known users

fix: remove skip button, add Anonymous{n} numbering for nameless users

Fix gr.BrowserState bug by replacing it with gr.State

Fix syntax error in app.py

Add ToS modal and Supabase REST analytics

Migrate from supabase python sdk to native requests

chore: migrate to supabase for analytics

Fix analytics IP and timeouts

Add backend analytics using counterapi.dev

Fix model loading by explicitly passing HF_TOKEN

Fix tokenizer crash by overriding extra_special_tokens

Fix dark mode credits visibility

Switch to QAT-Merged safetensors to fix inference bug

Fix CSS credits and NameError _analyze_essay_gpu

Fix ZeroGPU inference and CSS bugs

Hotfix: Replace Gradio CSS variables with explicit custom standard CSS classes for Credits

Hotfix: Use Gradio CSS variables for Credits section

UI Premium update and ZeroGPU bug fix

Architectural overhaul: Remove RAG, fix GPU bug, and update UI

Refactor: Remove RAG and Chat, Fix GPU bug, Update UI

Enable Gradio share link

Fix GPU word count bug and eradicate UI slop

Remove RAG and chat mode, update UI

Remove Chat mode, update to Gemma 4 QAT model

Update license to proprietary commercial license

Fix layout transition on empty input and hide reset button during load

Remove inputs block from custom Gradio theme to fix launch crash

Ignore compiled pycache and clean repository

Fix custom Gradio theme color assignment

Redesigned UI with premium layout and removed chatmode

Fix Gradio/Pydantic schema collision on launch

Fix inline gr.State instantiation bug in Gradio

Upgrade Viorra Space to v2.0: NumPy RAG, interactive chat, Gemma 4

Update parsing logic and prompts to support strict qualitative JSON and SOUL.md

Port architecture: Inject SOUL.md and implement <think> tag extraction in Gradio UI

chore: Sync UK rejection heuristic and update credits

Update disclaimer about Gradio UI

Clean Initial Commit for VIORRA 1.0.0