Require the API key on /v1 and default thinking off

Use LLAMA_API_KEY (not gateway's own API_KEY) to authenticate to llama-server

Forward the API key to llama-server on proxied requests

Fix llama-server failing to start: set LD_LIBRARY_PATH

Serve llama.cpp web UI on port 7860 via gateway proxy

Update gateway.py

Rename gateway (1).py to gateway.py

Delete gateway.py

Upload gateway (1).py

Upload 2 files

Update gateway.py

Update start.sh

Update gateway.py

Upload 4 files

Update Dockerfile

Create Dockerfile

Create README.md

initial commit