diff --git a/config/ai_config.env.template b/config/ai_config.env.template
index 5c172a85..6778b9f9 100644
--- a/config/ai_config.env.template
+++ b/config/ai_config.env.template
@@ -1,5 +1,8 @@
 # Copy this template to config/ai_config.env and adjust the following variables:
+LOCAL_AI_SCHEME="http"
 LOCAL_AI_IP="192.168.111.1"
+LOCAL_AI_PORT="8080"
+LOCAL_AI_API_ENDPOINT="/v1"
 # EMBA checks if this model is available for chatting
 # if this model is not available EMBA can't proceed
 LOCAL_AI_MODEL="Qwen2.5-Coder-7B-Instruct-GGUF"
diff --git a/helpers/helpers_emba_defaults.sh b/helpers/helpers_emba_defaults.sh
index 21c6f67a..55e938f8 100755
--- a/helpers/helpers_emba_defaults.sh
+++ b/helpers/helpers_emba_defaults.sh
@@ -94,6 +94,9 @@ set_defaults() {
   # Todo: remove option 1 and 2 and GPT_QUESTION usage
   # Todo: rename all GPT to AI areas
   export AI_OPTION=0 # 0 -> off, 1-> unpaid plan, 2 -> no rate-limit, 3 -> LocalAI Server
+  export LOCAL_AI_SCHEME="http"
+  export LOCAL_AI_PORT="8080"
+  export LOCAL_AI_API_ENDPOINT="/v1"
   export GPT_QUESTION=""
   export MINIMUM_GPT_PRIO=1 # everything above this value gets checked
 
diff --git a/modules/Q03_localai_connector.sh b/modules/Q03_localai_connector.sh
index 0cef0f1a..85cb5a7a 100755
--- a/modules/Q03_localai_connector.sh
+++ b/modules/Q03_localai_connector.sh
@@ -22,6 +22,14 @@
 : "${DEBUG:=0}"
 export DEBUG
 
+local_ai_api_url() {
+  local lAPI_PATH="${1#/}"
+  local lAPI_ENDPOINT="/${LOCAL_AI_API_ENDPOINT#/}"
+  lAPI_ENDPOINT="${lAPI_ENDPOINT%/}"
+
+  printf '%s://%s:%s%s/%s' "${LOCAL_AI_SCHEME}" "${LOCAL_AI_IP}" "${LOCAL_AI_PORT}" "${lAPI_ENDPOINT}" "${lAPI_PATH}"
+}
+
 Q03_localai_connector() {
   module_log_init "${FUNCNAME[0]}"
   export AI_RESULT_CNT=0
@@ -289,7 +297,7 @@ ask_localai() {
           '{model: $model, messages: [{role: "user", content: $prompt}], temperature: 0.2}')
 
         lstart_time=$(date +%s)
-        lHTTP_CODE=$(curl --connect-timeout 10 --max-time 600 -s http://"${LOCAL_AI_IP}":8080/v1/chat/completions \
+        lHTTP_CODE=$(curl --connect-timeout 10 --max-time 600 -s "$(local_ai_api_url "chat/completions")" \
           -H "Content-Type: application/json" \
           -d "${lJSON_PAYLOAD}" -o "${lAI_LOG_FILE}.json" --write-out "%{http_code}" || true)
 
@@ -416,7 +424,7 @@ identify_ai_model() {
   local lMODEL_LOCALAI=""
   local lCNT=1
   while [[ -z "${lMODEL_LOCALAI}" ]]; do
-    lMODEL_LOCALAI=$(curl --connect-timeout 10 --max-time 30 http://"${LOCAL_AI_IP}":8080/v1/models 2>/dev/null | jq -r .data[].id || true)
+    lMODEL_LOCALAI=$(curl --connect-timeout 10 --max-time 30 "$(local_ai_api_url "models")" 2>/dev/null | jq -r .data[].id || true)
     [[ -n "${lMODEL_LOCALAI}" ]] && break
     sleep 5
     lCNT=$((lCNT + 1))
diff --git a/tests/helpers/defaults_helpers.bats b/tests/helpers/defaults_helpers.bats
index a4361677..9b49e8e4 100644
--- a/tests/helpers/defaults_helpers.bats
+++ b/tests/helpers/defaults_helpers.bats
@@ -75,6 +75,13 @@ teardown() {
   [ "${SHELLCHECK}" -eq 1 ]
 }
 
+@test "set_defaults sets local connector endpoint defaults" {
+  set_defaults
+  [ "${LOCAL_AI_SCHEME}" = "http" ]
+  [ "${LOCAL_AI_PORT}" = "8080" ]
+  [ "${LOCAL_AI_API_ENDPOINT}" = "/v1" ]
+}
+
 @test "set_defaults sets SBOM_MAX_FILE_LOG" {
   set_defaults
   [ "${SBOM_MAX_FILE_LOG}" -eq 200 ]
diff --git a/tests/modules/localai_connector.bats b/tests/modules/localai_connector.bats
new file mode 100644
index 00000000..d2be6ca3
--- /dev/null
+++ b/tests/modules/localai_connector.bats
@@ -0,0 +1,51 @@
+# EMBA - EMBEDDED LINUX ANALYZER
+#
+# Copyright 2026-2026 Siemens Energy AG
+#
+# EMBA comes with ABSOLUTELY NO WARRANTY. This is free software, and you are
+# welcome to redistribute it under the terms of the GNU General Public License.
+# See LICENSE file for usage of this software.
+#
+# EMBA is licensed under GPLv3
+# SPDX-License-Identifier: GPL-3.0-only
+
+load ../setup.bash
+
+setup() {
+  setup_emba_test_env
+  # shellcheck source=helpers/helpers_emba_defaults.sh
+  source "${HELP_DIR}/helpers_emba_defaults.sh"
+  # shellcheck source=modules/Q03_localai_connector.sh
+  source "${MOD_DIR}/Q03_localai_connector.sh"
+  set_defaults
+  export LOCAL_AI_IP="192.0.2.1"
+}
+
+teardown() {
+  teardown_emba_test_env
+}
+
+@test "local_ai_api_url preserves the default endpoint" {
+  result="$(local_ai_api_url "models")"
+  [ "${result}" = "http://192.0.2.1:8080/v1/models" ]
+}
+
+@test "configuration template preserves the scan profile AI mode" {
+  export AI_OPTION=2
+  # shellcheck source=config/ai_config.env.template
+  source "${CONFIG_DIR}/ai_config.env.template"
+
+  [ "${AI_OPTION}" -eq 2 ]
+  result="$(local_ai_api_url "models")"
+  [ "${result}" = "http://192.168.111.1:8080/v1/models" ]
+}
+
+@test "local_ai_api_url uses configured endpoint components" {
+  result="$(LOCAL_AI_SCHEME="https" LOCAL_AI_PORT="9443" LOCAL_AI_API_ENDPOINT="/custom" local_ai_api_url "chat/completions")"
+  [ "${result}" = "https://192.0.2.1:9443/custom/chat/completions" ]
+}
+
+@test "local_ai_api_url normalizes endpoint separators" {
+  result="$(LOCAL_AI_API_ENDPOINT="custom/" local_ai_api_url "/models")"
+  [ "${result}" = "http://192.0.2.1:8080/custom/models" ]
+}