This is my script that I use for launching an instance of Claude Code using my local LLM, and it was written for an XDA article. Save it, edit the IP and port, and mark as executable to use.

lcc.sh

#!/usr/bin/env bash
# lcc - Local Claude Code launcher
# Points Claude Code at a local LLM served by llama.cpp on your GB10 device
#
# Usage:
#   lcc <modelname>           — launch Claude Code with the specified model
#   lcc <modelname> [args]    — pass additional arguments to claude
#   lcc                       — show help/launch with model if one is available
#
# Prerequisites:
#   - Claude Code installed
#   - LLM provider with Anthropic Messages API support
#   - jq (for model auto-detection)
#
#
# Configuration: change if needed:
LCC_HOST="${LCC_HOST:-192.168.1.179}" # This is my server IP, you should change to your most commonly used
LCC_PORT="${LCC_PORT:-8000}"
LCC_BASE_URL="http://${LCC_HOST}:${LCC_PORT}"
# ANSI color codes
RED='\033[0;31m'
GREEN='\033[0;32m'
YELLOW='\033[1;33m'
BLUE='\033[0;34m'
MAGENTA='\033[0;35m'
CYAN='\033[0;36m'
WHITE='\033[1;37m'
NC='\033[0m'
BOLD='\033[1m'
# No model argument, show help and check server
if [[ -z "$1" ]]; then
    echo -e "${BOLD}Local Claude Code${NC} (${CYAN}GB10 @ ${LCC_HOST}:${LCC_PORT}${NC})"
    echo ""
    # Check if server is reachable
    if curl -sf "${LCC_BASE_URL}/health" > /dev/null 2>&1; then
        echo -e "${GREEN} llama-server is running${NC}"
        # Try to get the loaded model name from /v1/models
        MODEL_INFO=$(curl -sf "${LCC_BASE_URL}/v1/models" 2>/dev/null)
        if [[ -n "$MODEL_INFO" ]]; then
            # Check for jq
            if ! command -v jq &>/dev/null; then
                echo ""
                echo -e "${YELLOW} jq not found; install it for model auto-detection${NC}"
            else
                echo ""
                echo -e "${WHITE}Loaded model(s):${NC}"
                MODEL_COUNT=$(echo "$MODEL_INFO" | jq '.data | length')
                # Store models in an array
                MODELS=()
                while IFS= read -r line; do
                    MODELS+=("$line")
                done < <(echo "$MODEL_INFO" | jq -r '.data[].id')
                # Print models with numbering
                for i in "${!MODELS[@]}"; do
                    if [[ "$MODEL_COUNT" -eq 1 ]]; then
                        echo -e "  ${GREEN}${NC} ${BOLD}${MODELS[$i]}${NC}"
                    else
                        echo -e "  ${WHITE}$((i+1)).${NC} ${MODELS[$i]}"
                    fi
                done
                # If exactly one model, use it
                if [[ "$MODEL_COUNT" -eq 1 ]]; then
                    MODEL="${MODELS[0]}"
                    echo ""
                    echo -e "${CYAN}Automatic model selection:${NC} Using ${BOLD}${MODEL}${NC}"
                    echo ""
                    echo -e "${GREEN} Launching Claude Code...${NC}"
                    export ANTHROPIC_BASE_URL="${LCC_BASE_URL}"
                    export ANTHROPIC_AUTH_TOKEN="local"
                    export ANTHROPIC_API_KEY=""
                    export CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC=1
                    exec claude --model "$MODEL" "$@"
                fi
            fi
        fi
    else
        echo -e "${RED} Cannot reach llama-server at ${LCC_BASE_URL}${NC}"
        echo ""
        echo ""
        echo -e "${WHITE}Make sure you're using an LLM provider that supports${NC}"
        echo -e "  ${CYAN}the Anthropic Messages API (/v1/messages endpoint)${NC}"
    fi
    echo ""
    echo -e "${WHITE}Examples:${NC}"
    echo -e "  ${GREEN}lcc qwen3-coder${NC}"
    echo -e "  ${GREEN}lcc my-model -p${NC} # pass extra flags to claude"
    echo ""
    echo -e "${WHITE}Override host/port:${NC}"
    echo -e "  ${CYAN}LCC_HOST=10.0.0.5 LCC_PORT=9090 lcc mymodel${NC}"
    exit 0
fi
# Launch Claude Code with the specified model
MODEL="$1"
shift  # remaining args pass through to claude
echo -e "${GREEN}Launching${NC} ${BOLD}Claude Code${NC} ${CYAN}${LCC_HOST}:${LCC_PORT}${NC} / ${MAGENTA}${BOLD}${MODEL}${NC}"
export ANTHROPIC_BASE_URL="${LCC_BASE_URL}"
export ANTHROPIC_AUTH_TOKEN="local"
export ANTHROPIC_API_KEY=""
export CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC=1
exec claude --model "$MODEL" "$@"