Remove hardcoded NVIDIA API key - now requires env var

Add NVIDIA API (GLM-4.7) as primary inference method with thinking mode

Fix HTTP 422: Only send fields expected by GameState model

Fix HTTP 422: Send full GameState to /api/infer endpoint

Add detailed logging to diagnose LLM inference failures

Fix extra closing div tag causing JSX build error

Fix fragment mismatch - move showMemoryLog inside container

Fix JSX syntax error in LLM output panel

Add live LLM output panel and AI decision factors display

Update inference.py

Update inference.py

Update inference.py

Update inference.py

Upload uv.lock

Update server/app.py

Update server/app.py

Update server/app.py

Update server/app.py

Update pyproject.toml

Upload pyproject.toml

Fix Nginx routing, resolve FastAPI route conflicts, and update inference.py requirements

Proxy /openenv/ through nginx

Replace HF token placeholder with safe example

Add OpenEnv API endpoints and fix reset/step/state support for Phase 1 checks

Sanitize ui/.env.example by removing hardcoded token

Restore ui folder in repository root

Add README configuration for HF Docker Space

Sanitize ui/.env.example by removing hardcoded token

Restore ui folder in repository root

Clean repository root for HF Spaces