Does anyone have tips for getting this to work with llamafile?

[TFWol]

llamafile supposedly has llama.cpp built-in, but when I run it with:
./llamafile -m qwen2.5-coder-3b-q8_0.gguf -l 192.168.1.123:8080 --trust "192.168.1.0/24" --server --v2 -ngl 999 --completion-mode
, point the extension at the server, and then start typing, I see errors spawning in the terminal:

llamafile/server/listen.cpp:41 server listen http://192.168.1.123:8080
llamafile/server/worker.cpp:143  warning: gpu mode disables pledge security
llamafile/server/client.cpp:679 192.168.1.100 POST /infill
llamafile/server/client.cpp:736 192.168.1.100 path not found: /zip/www/infill
llamafile/server/client.cpp:320 192.168.1.100 error 404 Not Found
llamafile/server/client.cpp:679 192.168.1.100 POST /infill
llamafile/server/client.cpp:736 192.168.1.100 path not found: /zip/www/infill
llamafile/server/client.cpp:320 192.168.1.100 error 404 Not Found
llamafile/server/client.cpp:679 192.168.1.100 POST /infill
llamafile/server/client.cpp:736 192.168.1.100 path not found: /zip/www/infill
llamafile/server/client.cpp:320 192.168.1.100 error 404 Not Found

Is there something I can tweak in the extension?