arktrace-llm-setup

name	arktrace-llm-setup
description	Configure the LLM provider for arktrace (OpenAI, local Ollama, or llama.cpp). Use when setting up a new environment or switching LLM providers.
license	Apache-2.0
compatibility	Requires .env file; optional local GPU for self-hosted models
metadata	{"repo":"arktrace"}

Set LLM_PROVIDER in .env to one of:

Value	Where it runs	Requires
openai	OpenAI API	OPENAI_API_KEY in .env
ollama	Local (Ollama)	ollama serve running on :11434
llama_cpp	Local (llama.cpp)	server running on LLM_BASE_URL

Ollama quick start

ollama serve                  # separate terminal
ollama pull llama3.2
echo 'LLM_PROVIDER=ollama' >> .env

See references/local-llm-setup.md for model selection, GPU config, and performance tuning.