Update startup.sh

Update model_manager.py

Update Dockerfile

Update app.py

Update Dockerfile

Update app.py

Update app.py

Update app.py

Update README.md

Update Dockerfile

Update startup.sh

Update model_manager.py

Update app.py

Update app.py

Upload app.py

Delete app.py

Update app.py

Update app.py

Update Dockerfile

Update requirements.txt

Update app.py

Update model_manager.py

Update model_manager.py

Update startup.sh

Create startup.sh

Create model_manager.py

Create app.py

Update requirements.txt

Update README.md

Delete app.py

Update app.py

Update Dockerfile

🔥 GOD MODE ACTIVATED: Subprocess monitoring, Context Guardrails, and No-MMAP Turbo Boost

🚦 TRAFFIC JAM HACK: Added Parallel Processing & Invisible Keep-Alive Heartbeats

🧠 Expand Context Window to 8192 tokens for long documents

🕷️ OMNI-SCRAPER: Added YouTube transcript, Image download & Article reading from URLs

🚀 FINAL PRODUCTION BUILD: Added Auth, CORS, and Smart Web Scraper

🔧 Fix llama-server flag: specify --flash-attn on

🚀 PERFECT OPTIMIZATION: KV Cache 8-bit Quantization & Flash Attn for max CPU speed

🔧 Remove unsupported --image-max-pixels flag

🔧 Fix OOM crash: reduce memory usage and image bounds

🚀 Deploy Universal Engine: aiohttp, PDF, Video & multimodal processing

🔥 FINAL FIX: remove streaming completely

Fix proxy JSON parsing and streaming headers

🔥 FINAL ULTRA SYSTEM

🔥 SPEED OPTIMIZATION V3

Final performance tuning: threads 2/4 + mlock

Use Python script for reliable model download

Switch to reliable huggingface-cli for model download

Fix flash-attn syntax and revert to stable llama.cpp