diff --git a/.github/ISSUE_TEMPLATE/bug_report.md b/.github/ISSUE_TEMPLATE/bug_report.md
deleted file mode 100644
index 837bdb720..000000000
--- a/.github/ISSUE_TEMPLATE/bug_report.md
+++ /dev/null
@@ -1,38 +0,0 @@
----
-name: Bug report
-about: Create a report to help us improve
-title: 'bug: [DESCRIPTION]'
-labels: 'type: bug'
-assignees: ''
-
----
-
-**Describe the bug**
-A clear and concise description of what the bug is.
-
-**To Reproduce**
-Steps to reproduce the behavior:
-1. Go to '...'
-2. Click on '....'
-3. Scroll down to '....'
-4. See error
-
-**Expected behavior**
-A clear and concise description of what you expected to happen.
-
-**Screenshots**
-If applicable, add screenshots to help explain your problem.
-
-**Desktop (please complete the following information):**
- - OS: [e.g. iOS]
- - Browser [e.g. chrome, safari]
- - Version [e.g. 22]
-
-**Smartphone (please complete the following information):**
- - Device: [e.g. iPhone6]
- - OS: [e.g. iOS8.1]
- - Browser [e.g. stock browser, safari]
- - Version [e.g. 22]
-
-**Additional context**
-Add any other context about the problem here.
diff --git a/.github/ISSUE_TEMPLATE/bug_report.yml b/.github/ISSUE_TEMPLATE/bug_report.yml
new file mode 100644
index 000000000..d0182a83a
--- /dev/null
+++ b/.github/ISSUE_TEMPLATE/bug_report.yml
@@ -0,0 +1,50 @@
+name: "\U0001F41B Bug Report"
+description: "If something isn't working as expected \U0001F914"
+labels: [ "type: bug" ]
+title: 'bug: [DESCRIPTION]'
+
+body:
+  - type: input
+    validations:
+      required: true
+    attributes:
+      label: "Cortex version"
+      description: "**Tip:** The version is in the app's bottom right corner"
+
+  - type: textarea
+    validations:
+      required: true
+    attributes:
+      label: "Describe the Bug"
+      description: "A clear & concise description of the bug"
+
+  - type: textarea
+    attributes:
+      label: "Steps to Reproduce"
+      description: |
+        Please list out steps to reproduce the issue
+      placeholder: |
+        1. Go to '...'
+        2. Click on '...'
+
+  - type: textarea
+    attributes:
+      label: "Screenshots / Logs"
+      description: |
+        You can find logs in: ~/cortex/logs
+
+  - type: checkboxes
+    attributes:
+      label: "What is your OS?"
+      options:
+        - label: MacOS
+        - label: Windows
+        - label: Linux
+
+  - type: checkboxes
+    attributes:
+      label: "What engine are you running?"
+      options:
+        - label: cortex.llamacpp (default)
+        - label: cortex.tensorrt-llm (Nvidia GPUs)
+        - label: cortex.onnx (NPUs, DirectML)
\ No newline at end of file
diff --git a/.github/ISSUE_TEMPLATE/config.yml b/.github/ISSUE_TEMPLATE/config.yml
new file mode 100644
index 000000000..56e11b10a
--- /dev/null
+++ b/.github/ISSUE_TEMPLATE/config.yml
@@ -0,0 +1,7 @@
+## To encourage contributors to use issue templates, we don't allow blank issues
+blank_issues_enabled: true
+
+contact_links:
+  - name: "\1F4AC Cortex Discussions"
+    url: "https://github.com/orgs/janhq/discussions/categories/q-a"
+    about: "Get help, discuss features & roadmap, and share your projects"
\ No newline at end of file
diff --git a/.github/ISSUE_TEMPLATE/discussion-thread.md b/.github/ISSUE_TEMPLATE/discussion-thread.md
deleted file mode 100644
index 09e52ae68..000000000
--- a/.github/ISSUE_TEMPLATE/discussion-thread.md
+++ /dev/null
@@ -1,14 +0,0 @@
----
-name: Discussion thread
-about: Start an open ended discussion
-title: 'Discussion: [TOPIC HERE]'
-labels: ''
-assignees: ''
-
----
-
-**Motivation**
-
-**Discussion**
-
-**Resources**
diff --git a/.github/ISSUE_TEMPLATE/epic-request.md b/.github/ISSUE_TEMPLATE/epic-request.md
index bfad8e5f8..486ae90b6 100644
--- a/.github/ISSUE_TEMPLATE/epic-request.md
+++ b/.github/ISSUE_TEMPLATE/epic-request.md
@@ -8,13 +8,10 @@ assignees: ''
 ---
 
 **Problem**
-A clear and concise description of what the problem is. Ex. I'm always frustrated when [...]
 
-**Success Criteria**
-A clear and concise description of what you want to happen.
 
-**Sub Issues**
+**Success Criteria**
 - 
 
-**Additional context**
-Add any other context or screenshots about the epic request here.
+**Tasklist**
+- [ ] 
diff --git a/.github/ISSUE_TEMPLATE/feature_request.md b/.github/ISSUE_TEMPLATE/feature_request.md
deleted file mode 100644
index 26f586bd0..000000000
--- a/.github/ISSUE_TEMPLATE/feature_request.md
+++ /dev/null
@@ -1,17 +0,0 @@
----
-name: Feature request
-about: Suggest an idea for this project
-title: 'feat: [DESCRIPTION]'
-labels: 'type: feature request'
-assignees: ''
-
----
-
-**Problem**
-A clear and concise description of what the problem is. Ex. I'm always frustrated when [...]
-
-**Success Criteria**
-A clear and concise description of what you want to happen.
-
-**Additional context**
-Add any other context or screenshots about the feature request here.
diff --git a/.github/ISSUE_TEMPLATE/feature_request.yml b/.github/ISSUE_TEMPLATE/feature_request.yml
new file mode 100644
index 000000000..1d267e500
--- /dev/null
+++ b/.github/ISSUE_TEMPLATE/feature_request.yml
@@ -0,0 +1,19 @@
+name: "\U0001F680 Feature Request"
+description: "Suggest an idea for this project \U0001F63B!"
+title: 'idea: [DESCRIPTION]'
+body:
+  - type: textarea
+    validations:
+      required: true
+    attributes:
+      label: "Problem Statement"
+      description: "Describe the problem you're facing"
+      placeholder: |
+        I'm always frustrated when ...
+
+  - type: textarea
+    validations:
+      required: true
+    attributes:
+      label: "Feature Idea"
+      description: "Describe what you want instead. Examples are welcome!"
\ No newline at end of file
diff --git a/.github/ISSUE_TEMPLATE/model_request.yml b/.github/ISSUE_TEMPLATE/model_request.yml
new file mode 100644
index 000000000..c424de8fc
--- /dev/null
+++ b/.github/ISSUE_TEMPLATE/model_request.yml
@@ -0,0 +1,21 @@
+name: "\U0001F929 Model Request"
+description: "Request a new model to be compiled"
+title: 'feat: [DESCRIPTION]'
+labels: 'type: model request'
+body:
+  - type: markdown
+    attributes:
+      value: "**Tip:** Download any model with `cortex pull HUGGINGFACE_MODEL_ID`. Use this form for unsupported models only."
+  - type: textarea
+    validations:
+      required: true
+    attributes:
+      label: "Model Requests"
+      description: "If applicable, include the source URL, licenses, and any other relevant information"
+  - type: checkboxes
+    attributes:
+      label: "Which formats?"
+      options:
+        - label: GGUF (llama.cpp)
+        - label: TensorRT (TensorRT-LLM)
+        - label: ONNX (Onnx Runtime)
diff --git a/engine/CMakeLists.txt b/engine/CMakeLists.txt
index 3be4705da..58661c508 100644
--- a/engine/CMakeLists.txt
+++ b/engine/CMakeLists.txt
@@ -86,7 +86,7 @@ if(DEFINED CMAKE_JS_INC)
   # define NPI_VERSION
   add_compile_definitions(NAPI_VERSION=8)
 endif()
-
+ 
 add_compile_definitions(CORTEX_VARIANT="${CORTEX_VARIANT}")
 add_compile_definitions(CORTEX_CPP_VERSION="${CORTEX_CPP_VERSION}")
 add_compile_definitions(CORTEX_CONFIG_FILE_PATH="${CORTEX_CONFIG_FILE_PATH}")
@@ -115,6 +115,7 @@ if(DEFINED CMAKE_JS_INC)
 
   add_library(${PROJECT_NAME} SHARED addon.cc
     ${CMAKE_CURRENT_SOURCE_DIR}/utils/cpuid/cpu_info.cc
+    ${CMAKE_CURRENT_SOURCE_DIR}/utils/file_logger.cc
     ${CMAKE_JS_SRC}
   )
 
@@ -131,6 +132,7 @@ if(DEFINED CMAKE_JS_INC)
 else() # Official build
   add_executable(${PROJECT_NAME} main.cc
     ${CMAKE_CURRENT_SOURCE_DIR}/utils/cpuid/cpu_info.cc
+    ${CMAKE_CURRENT_SOURCE_DIR}/utils/file_logger.cc
   )
 endif()
 
diff --git a/engine/commands/cortex_upd_cmd.cc b/engine/commands/cortex_upd_cmd.cc
index 9027203e7..a4343e55c 100644
--- a/engine/commands/cortex_upd_cmd.cc
+++ b/engine/commands/cortex_upd_cmd.cc
@@ -4,6 +4,7 @@
 #include "cortex_upd_cmd.h"
 #include "httplib.h"
 #include "nlohmann/json.hpp"
+#include "server_stop_cmd.h"
 #include "services/download_service.h"
 #include "utils/archive_utils.h"
 #include "utils/file_manager_utils.h"
@@ -12,21 +13,38 @@
 
 namespace commands {
 
-namespace {
-const std::string kCortexBinary = "cortex-cpp";
-}
-
 CortexUpdCmd::CortexUpdCmd() {}
 
 void CortexUpdCmd::Exec(std::string v) {
-  // TODO(sang) stop server if it is running
+  {
+    auto config = file_manager_utils::GetCortexConfig();
+    httplib::Client cli(config.apiServerHost + ":" + config.apiServerPort);
+    auto res = cli.Get("/health/healthz");
+    if (res) {
+      CLI_LOG("Server is running. Stopping server before updating!");
+      commands::ServerStopCmd ssc(config.apiServerHost,
+                                  std::stoi(config.apiServerPort));
+      ssc.Exec();
+    }
+  }
+  if (CORTEX_VARIANT == file_manager_utils::kNightlyVariant) {
+    if (!GetNightly(v))
+      return;
+  } else {
+    if (!GetStableAndBeta(v))
+      return;
+  }
+  CLI_LOG("Update cortex sucessfully");
+}
+
+bool CortexUpdCmd::GetStableAndBeta(const std::string& v) {
   // Check if the architecture and OS are supported
   auto system_info = system_info_utils::GetSystemInfo();
   if (system_info.arch == system_info_utils::kUnsupported ||
       system_info.os == system_info_utils::kUnsupported) {
     CTL_ERR("Unsupported OS or architecture: " << system_info.os << ", "
                                                << system_info.arch);
-    return;
+    return false;
   }
   CTL_INF("OS: " << system_info.os << ", Arch: " << system_info.arch);
 
@@ -50,7 +68,7 @@ void CortexUpdCmd::Exec(std::string v) {
         std::string matched_variant = "";
         for (auto& asset : assets) {
           auto asset_name = asset["name"].get<std::string>();
-          if (asset_name.find("cortex-cpp") != std::string::npos &&
+          if (asset_name.find(kCortexBinary) != std::string::npos &&
               asset_name.find(os_arch) != std::string::npos) {
             matched_variant = asset_name;
             break;
@@ -59,7 +77,7 @@ void CortexUpdCmd::Exec(std::string v) {
         }
         if (matched_variant.empty()) {
           CTL_ERR("No variant found for " << os_arch);
-          return;
+          return false;
         }
         CTL_INF("Matched variant: " << matched_variant);
 
@@ -99,14 +117,6 @@ void CortexUpdCmd::Exec(std::string v) {
                   archive_utils::ExtractArchive(download_path.string(),
                                                 extract_path.string());
 
-                  // remove the downloaded file
-                  // TODO(any) Could not delete file on Windows because it is currently hold by httplib(?)
-                  // Not sure about other platforms
-                  try {
-                    std::filesystem::remove(absolute_path);
-                  } catch (const std::exception& e) {
-                    CTL_WRN("Could not delete file: " << e.what());
-                  }
                   CTL_INF("Finished!");
                 });
             break;
@@ -114,58 +124,75 @@ void CortexUpdCmd::Exec(std::string v) {
         }
       } catch (const nlohmann::json::parse_error& e) {
         std::cerr << "JSON parse error: " << e.what() << std::endl;
-        return;
+        return false;
       }
     } else {
       CTL_ERR("HTTP error: " << res->status);
-      return;
+      return false;
     }
   } else {
     auto err = res.error();
     CTL_ERR("HTTP error: " << httplib::to_string(err));
-    return;
+    return false;
   }
-#if defined(_WIN32)
-  auto executable_path = file_manager_utils::GetExecutableFolderContainerPath();
-  auto temp = executable_path / "cortex_tmp.exe";
-  remove(temp.string().c_str());  // ignore return code
-
-  auto src =
-      executable_path / "cortex" / kCortexBinary / (kCortexBinary + ".exe");
-  auto dst = executable_path / (kCortexBinary + ".exe");
-  // Rename
-  rename(dst.string().c_str(), temp.string().c_str());
-  // Update
-  CopyFile(const_cast<char*>(src.string().c_str()),
-           const_cast<char*>(dst.string().c_str()), false);
-  auto download_folder = executable_path / "cortex";
-  remove(download_folder);
-  remove(temp.string().c_str());
-#else
+
+  // Replace binary file
   auto executable_path = file_manager_utils::GetExecutableFolderContainerPath();
-  auto temp = executable_path / "cortex_tmp";
-  auto src = executable_path / "cortex" / kCortexBinary / kCortexBinary;
-  auto dst = executable_path / kCortexBinary;
-  if (std::rename(dst.string().c_str(), temp.string().c_str())) {
-    CTL_ERR("Failed to rename from " << dst.string() << " to "
-                                     << temp.string());
-    return;
-  }
-  try {
-    std::filesystem::copy_file(
-        src, dst, std::filesystem::copy_options::overwrite_existing);
-    std::filesystem::permissions(dst, std::filesystem::perms::owner_all |
-                                          std::filesystem::perms::group_all |
-                                          std::filesystem::perms::others_read |
-                                          std::filesystem::perms::others_exec);
-    std::filesystem::remove(temp);
-    auto download_folder = executable_path / "cortex/";
-    std::filesystem::remove_all(download_folder);
-  } catch (const std::exception& e) {
-    CTL_WRN("Something wrong happened: " << e.what());
-    return;
+  auto src = executable_path / "cortex" / kCortexBinary / GetCortexBinary();
+  auto dst = executable_path / GetCortexBinary();
+  return ReplaceBinaryInflight(src, dst);
+}
+
+bool CortexUpdCmd::GetNightly(const std::string& v) {
+  // Check if the architecture and OS are supported
+  auto system_info = system_info_utils::GetSystemInfo();
+  if (system_info.arch == system_info_utils::kUnsupported ||
+      system_info.os == system_info_utils::kUnsupported) {
+    CTL_ERR("Unsupported OS or architecture: " << system_info.os << ", "
+                                               << system_info.arch);
+    return false;
   }
-#endif
-  CLI_LOG("Update cortex sucessfully");
+  CTL_INF("OS: " << system_info.os << ", Arch: " << system_info.arch);
+
+  // Download file
+  std::string version = v.empty() ? "latest" : std::move(v);
+  std::ostringstream release_path;
+  release_path << "cortex/" << version << "/" << system_info.os << "-"
+               << system_info.arch << "/" << kNightlyFileName;
+  CTL_INF("Engine release path: " << kNightlyHost << release_path.str());
+
+  auto download_task = DownloadTask{.id = "cortex",
+                                    .type = DownloadType::Cortex,
+                                    .error = std::nullopt,
+                                    .items = {DownloadItem{
+                                        .id = "cortex",
+                                        .host = kNightlyHost,
+                                        .fileName = kNightlyFileName,
+                                        .type = DownloadType::Cortex,
+                                        .path = release_path.str(),
+                                    }}};
+
+  DownloadService download_service;
+  download_service.AddDownloadTask(
+      download_task, [this](const std::string& absolute_path, bool unused) {
+        // try to unzip the downloaded file
+        std::filesystem::path download_path{absolute_path};
+        CTL_INF("Downloaded engine path: " << download_path.string());
+
+        std::filesystem::path extract_path =
+            download_path.parent_path().parent_path();
+
+        archive_utils::ExtractArchive(download_path.string(),
+                                      extract_path.string());
+
+        CTL_INF("Finished!");
+      });
+
+  // Replace binay file
+  auto executable_path = file_manager_utils::GetExecutableFolderContainerPath();
+  auto src = executable_path / "cortex" / GetCortexBinary();
+  auto dst = executable_path / GetCortexBinary();
+  return ReplaceBinaryInflight(src, dst);
 }
+
 }  // namespace commands
\ No newline at end of file
diff --git a/engine/commands/cortex_upd_cmd.h b/engine/commands/cortex_upd_cmd.h
index 2606dbbd9..396c6cdaf 100644
--- a/engine/commands/cortex_upd_cmd.h
+++ b/engine/commands/cortex_upd_cmd.h
@@ -1,13 +1,124 @@
 #pragma once
-#include <string>
 #include <optional>
+#include <string>
+
+#include "httplib.h"
+#include "nlohmann/json.hpp"
+#include "utils/file_manager_utils.h"
+#include "utils/logging_utils.h"
 
 namespace commands {
+#ifndef CORTEX_VARIANT
+#define CORTEX_VARIANT file_manager_utils::kProdVariant
+#endif
+constexpr const auto kNightlyHost = "https://delta.jan.ai";
+constexpr const auto kNightlyFileName = "cortex-nightly.tar.gz";
+const std::string kCortexBinary = "cortex";
+
+inline std::string GetCortexBinary() {
+#if defined(_WIN32)
+  constexpr const bool has_exe = true;
+#else
+  constexpr const bool has_exe = false;
+#endif
+  if (CORTEX_VARIANT == file_manager_utils::kNightlyVariant) {
+    return has_exe ? kCortexBinary + "-nightly.exe"
+                   : kCortexBinary + "-nightly";
+  } else if (CORTEX_VARIANT == file_manager_utils::kBetaVariant) {
+    return has_exe ? kCortexBinary + "-beta.exe" : kCortexBinary + "-beta";
+  } else {
+    return has_exe ? kCortexBinary + ".exe" : kCortexBinary;
+  }
+}
+
+inline std::string GetHostName() {
+  if (CORTEX_VARIANT == file_manager_utils::kNightlyVariant) {
+    return "https://delta.jan.ai";
+  } else {
+    return "https://api.github.com";
+  }
+}
+
+inline std::string GetReleasePath() {
+  if (CORTEX_VARIANT == file_manager_utils::kNightlyVariant) {
+    return "/cortex/latest/version.json";
+  } else {
+    return "/repos/janhq/cortex.cpp/releases/latest";
+  }
+}
+
+inline void CheckNewUpdate() {
+  auto host_name = GetHostName();
+  auto release_path = GetReleasePath();
+  CTL_INF("Engine release path: " << host_name << release_path);
 
-class CortexUpdCmd{
+  httplib::Client cli(host_name);
+  if (auto res = cli.Get(release_path)) {
+    if (res->status == httplib::StatusCode::OK_200) {
+      try {
+        auto json_res = nlohmann::json::parse(res->body);
+        std::string latest_version = json_res["tag_name"].get<std::string>();
+        std::string current_version = CORTEX_CPP_VERSION;
+        if (current_version != latest_version) {
+          CLI_LOG("\nA new release of cortex is available: "
+                  << current_version << " -> " << latest_version);
+          CLI_LOG("To upgrade, run: cortex update");
+          // CLI_LOG(json_res["html_url"].get<std::string>());
+        }
+      } catch (const nlohmann::json::parse_error& e) {
+        CTL_INF("JSON parse error: " << e.what());
+      }
+    } else {
+      CTL_INF("HTTP error: " << res->status);
+    }
+  } else {
+    auto err = res.error();
+    CTL_INF("HTTP error: " << httplib::to_string(err));
+  }
+}
+
+inline bool ReplaceBinaryInflight(const std::filesystem::path& src,
+                                  const std::filesystem::path& dst) {
+  if (src == dst) {
+    // Already has the newest
+    return true;
+  }
+  std::filesystem::path temp = std::filesystem::temp_directory_path() / "cortex_temp";
+
+  try {
+    if (std::filesystem::exists(temp)) {
+      std::filesystem::remove(temp);
+    }
+
+    std::rename(dst.string().c_str(), temp.string().c_str());
+    std::filesystem::copy_file(
+        src, dst, std::filesystem::copy_options::overwrite_existing);
+    std::filesystem::permissions(dst, std::filesystem::perms::owner_all |
+                                          std::filesystem::perms::group_all |
+                                          std::filesystem::perms::others_read |
+                                          std::filesystem::perms::others_exec);
+    auto download_folder = src.parent_path();
+    std::filesystem::remove_all(download_folder);
+  } catch (const std::exception& e) {
+    CTL_ERR("Something wrong happened: " << e.what());
+    if (std::filesystem::exists(temp)) {
+      std::rename(temp.string().c_str(), dst.string().c_str());
+      CLI_LOG("Restored binary file");
+    }
+    return false;
+  }
+
+  return true;
+}
+
+class CortexUpdCmd {
  public:
   CortexUpdCmd();
   void Exec(std::string version);
+
+ private:
+  bool GetStableAndBeta(const std::string& v);
+  bool GetNightly(const std::string& v);
 };
 
 }  // namespace commands
\ No newline at end of file
diff --git a/engine/commands/engine_init_cmd.cc b/engine/commands/engine_init_cmd.cc
index b3ab8bae5..7b0153f69 100644
--- a/engine/commands/engine_init_cmd.cc
+++ b/engine/commands/engine_init_cmd.cc
@@ -75,7 +75,9 @@ bool EngineInitCmd::Exec() const {
           matched_variant = engine_matcher_utils::ValidateOnnx(
               variants, system_info.os, system_info.arch);
         } else if (engineName_ == "cortex.llamacpp") {
-          auto suitable_avx = engine_matcher_utils::GetSuitableAvxVariant();
+          cortex::cpuid::CpuInfo cpu_info;
+          auto suitable_avx =
+              engine_matcher_utils::GetSuitableAvxVariant(cpu_info);
           matched_variant = engine_matcher_utils::Validate(
               variants, system_info.os, system_info.arch, suitable_avx,
               cuda_driver_version);
diff --git a/engine/controllers/command_line_parser.cc b/engine/controllers/command_line_parser.cc
index b8ad72f58..987aa693a 100644
--- a/engine/controllers/command_line_parser.cc
+++ b/engine/controllers/command_line_parser.cc
@@ -16,7 +16,6 @@
 #include "config/yaml_config.h"
 #include "httplib.h"
 #include "services/engine_service.h"
-#include "utils/cortex_utils.h"
 #include "utils/file_manager_utils.h"
 #include "utils/logging_utils.h"
 
@@ -24,7 +23,7 @@ CommandLineParser::CommandLineParser()
     : app_("Cortex.cpp CLI"), engine_service_{EngineService()} {}
 
 bool CommandLineParser::SetupCommand(int argc, char** argv) {
-
+  auto config = file_manager_utils::GetCortexConfig();
   std::string model_id;
 
   // Models group commands
@@ -34,7 +33,7 @@ bool CommandLineParser::SetupCommand(int argc, char** argv) {
 
     auto start_cmd = models_cmd->add_subcommand("start", "Start a model by ID");
     start_cmd->add_option("model_id", model_id, "");
-    start_cmd->callback([&model_id]() {
+    start_cmd->callback([&model_id, &config]() {
       commands::CmdInfo ci(model_id);
       std::string model_file =
           ci.branch == "main" ? ci.model_name : ci.model_name + "-" + ci.branch;
@@ -42,7 +41,8 @@ bool CommandLineParser::SetupCommand(int argc, char** argv) {
       yaml_handler.ModelConfigFromFile(
           file_manager_utils::GetModelsContainerPath().string() + "/" +
           model_file + ".yaml");
-      commands::ModelStartCmd msc("127.0.0.1", 3928,
+      commands::ModelStartCmd msc(config.apiServerHost,
+                                  std::stoi(config.apiServerPort),
                                   yaml_handler.GetModelConfig());
       msc.Exec();
     });
@@ -50,7 +50,7 @@ bool CommandLineParser::SetupCommand(int argc, char** argv) {
     auto stop_model_cmd =
         models_cmd->add_subcommand("stop", "Stop a model by ID");
     stop_model_cmd->add_option("model_id", model_id, "");
-    stop_model_cmd->callback([&model_id]() {
+    stop_model_cmd->callback([&model_id, &config]() {
       commands::CmdInfo ci(model_id);
       std::string model_file =
           ci.branch == "main" ? ci.model_name : ci.model_name + "-" + ci.branch;
@@ -58,7 +58,8 @@ bool CommandLineParser::SetupCommand(int argc, char** argv) {
       yaml_handler.ModelConfigFromFile(
           file_manager_utils::GetModelsContainerPath().string() + "/" +
           model_file + ".yaml");
-      commands::ModelStopCmd smc("127.0.0.1", 3928,
+      commands::ModelStopCmd smc(config.apiServerHost,
+                                 std::stoi(config.apiServerPort),
                                  yaml_handler.GetModelConfig());
       smc.Exec();
     });
@@ -105,7 +106,7 @@ bool CommandLineParser::SetupCommand(int argc, char** argv) {
     chat_cmd->add_option("model_id", model_id, "");
     chat_cmd->add_option("-m,--message", msg, "Message to chat with model");
 
-    chat_cmd->callback([&model_id, &msg] {
+    chat_cmd->callback([&model_id, &msg, &config] {
       commands::CmdInfo ci(model_id);
       std::string model_file =
           ci.branch == "main" ? ci.model_name : ci.model_name + "-" + ci.branch;
@@ -113,7 +114,9 @@ bool CommandLineParser::SetupCommand(int argc, char** argv) {
       yaml_handler.ModelConfigFromFile(
           file_manager_utils::GetModelsContainerPath().string() + "/" +
           model_file + ".yaml");
-      commands::ChatCmd cc("127.0.0.1", 3928, yaml_handler.GetModelConfig());
+      commands::ChatCmd cc(config.apiServerHost,
+                           std::stoi(config.apiServerPort),
+                           yaml_handler.GetModelConfig());
       cc.Exec(msg);
     });
   }
@@ -135,9 +138,19 @@ bool CommandLineParser::SetupCommand(int argc, char** argv) {
       command.Exec();
     });
 
+    auto install_cmd = engines_cmd->add_subcommand("install", "Install engine");
+    install_cmd->callback([] { CLI_LOG("Engine name can't be empty!"); });
+    for (auto& engine : engine_service_.kSupportEngines) {
+      std::string engine_name{engine};
+      EngineInstall(install_cmd, engine_name, version);
+    }
+
+    auto uninstall_cmd =
+        engines_cmd->add_subcommand("uninstall", "Uninstall engine");
+    uninstall_cmd->callback([] { CLI_LOG("Engine name can't be empty!"); });
     for (auto& engine : engine_service_.kSupportEngines) {
       std::string engine_name{engine};
-      EngineManagement(engines_cmd, engine_name, version);
+      EngineUninstall(uninstall_cmd, engine_name);
     }
 
     EngineGet(engines_cmd);
@@ -149,17 +162,18 @@ bool CommandLineParser::SetupCommand(int argc, char** argv) {
         app_.add_subcommand("run", "Shortcut to start a model and chat");
     std::string model_id;
     run_cmd->add_option("model_id", model_id, "");
-    run_cmd->callback([&model_id] {
-      commands::RunCmd rc("127.0.0.1", 3928, model_id);
+    run_cmd->callback([&model_id, &config] {
+      commands::RunCmd rc(config.apiServerHost, std::stoi(config.apiServerPort),
+                          model_id);
       rc.Exec();
     });
   }
 
   auto stop_cmd = app_.add_subcommand("stop", "Stop the API server");
 
-  stop_cmd->callback([] {
-    // TODO get info from config file
-    commands::ServerStopCmd ssc("127.0.0.1", 3928);
+  stop_cmd->callback([&config] {
+    commands::ServerStopCmd ssc(config.apiServerHost,
+                                std::stoi(config.apiServerPort));
     ssc.Exec();
   });
 
@@ -193,59 +207,29 @@ bool CommandLineParser::SetupCommand(int argc, char** argv) {
   // Check new update, only check for stable release for now
 #ifdef CORTEX_CPP_VERSION
   if (check_update) {
-    constexpr auto github_host = "https://api.github.com";
-    std::ostringstream release_path;
-    release_path << "/repos/janhq/cortex.cpp/releases/latest";
-    CTL_INF("Engine release path: " << github_host << release_path.str());
-
-    httplib::Client cli(github_host);
-    if (auto res = cli.Get(release_path.str())) {
-      if (res->status == httplib::StatusCode::OK_200) {
-        try {
-          auto json_res = nlohmann::json::parse(res->body);
-          std::string latest_version = json_res["tag_name"].get<std::string>();
-          std::string current_version = CORTEX_CPP_VERSION;
-          if (current_version != latest_version) {
-            CLI_LOG("\nA new release of cortex is available: "
-                    << current_version << " -> " << latest_version);
-            CLI_LOG("To upgrade, run: cortex update");
-            CLI_LOG(json_res["html_url"].get<std::string>());
-          }
-        } catch (const nlohmann::json::parse_error& e) {
-          CTL_INF("JSON parse error: " << e.what());
-        }
-      } else {
-        CTL_INF("HTTP error: " << res->status);
-      }
-    } else {
-      auto err = res.error();
-      CTL_INF("HTTP error: " << httplib::to_string(err));
-    }
+    commands::CheckNewUpdate();
   }
 #endif
 
   return true;
 }
 
-void CommandLineParser::EngineManagement(CLI::App* parent,
-                                         const std::string& engine_name,
-                                         std::string& version) {
-  auto engine_cmd =
-      parent->add_subcommand(engine_name, "Manage " + engine_name + " engine");
-
-  auto install_cmd = engine_cmd->add_subcommand(
-      "install", "Install " + engine_name + " engine");
-  install_cmd->add_option("-v, --version", version,
-                          "Engine version. Default will be latest");
+void CommandLineParser::EngineInstall(CLI::App* parent,
+                                      const std::string& engine_name,
+                                      std::string& version) {
+  auto install_engine_cmd = parent->add_subcommand(engine_name, "");
 
-  install_cmd->callback([engine_name, &version] {
+  install_engine_cmd->callback([&] {
     commands::EngineInitCmd eic(engine_name, version);
     eic.Exec();
   });
+}
+
+void CommandLineParser::EngineUninstall(CLI::App* parent,
+                                        const std::string& engine_name) {
+  auto uninstall_engine_cmd = parent->add_subcommand(engine_name, "");
 
-  auto uninstall_desc{"Uninstall " + engine_name + " engine"};
-  auto uninstall_cmd = engine_cmd->add_subcommand("uninstall", uninstall_desc);
-  uninstall_cmd->callback([engine_name] {
+  uninstall_engine_cmd->callback([&] {
     commands::EngineUninstallCmd cmd(engine_name);
     cmd.Exec();
   });
diff --git a/engine/controllers/command_line_parser.h b/engine/controllers/command_line_parser.h
index 0fa89e241..e4a2f47c5 100644
--- a/engine/controllers/command_line_parser.h
+++ b/engine/controllers/command_line_parser.h
@@ -9,8 +9,10 @@ class CommandLineParser {
   bool SetupCommand(int argc, char** argv);
 
  private:
-  void EngineManagement(CLI::App* parent, const std::string& engine_name,
-                        std::string& version);
+  void EngineInstall(CLI::App* parent, const std::string& engine_name,
+                     std::string& version);
+
+  void EngineUninstall(CLI::App* parent, const std::string& engine_name);
 
   void EngineGet(CLI::App* parent);
 
diff --git a/engine/main.cc b/engine/main.cc
index 5c83d1103..1450d887c 100644
--- a/engine/main.cc
+++ b/engine/main.cc
@@ -6,6 +6,7 @@
 #include "utils/archive_utils.h"
 #include "utils/cortex_utils.h"
 #include "utils/dylib.h"
+#include "utils/file_logger.h"
 #include "utils/file_manager_utils.h"
 #include "utils/logging_utils.h"
 
@@ -26,19 +27,21 @@
 
 void RunServer() {
   auto config = file_manager_utils::GetCortexConfig();
-  LOG_INFO << "Host: " << config.host << " Port: " << config.port << "\n";
+  LOG_INFO << "Host: " << config.apiServerHost << " Port: " << config.apiServerPort << "\n";
 
   // Create logs/ folder and setup log to file
-  std::filesystem::create_directory(cortex_utils::logs_folder);
-  trantor::AsyncFileLogger asyncFileLogger;
-  asyncFileLogger.setFileName(cortex_utils::logs_base_name);
+  std::filesystem::create_directory(config.logFolderPath + "/" +
+                                    cortex_utils::logs_folder);
+  trantor::FileLogger asyncFileLogger;
+  asyncFileLogger.setFileName(config.logFolderPath + "/" +
+                              cortex_utils::logs_base_name);
+  asyncFileLogger.setMaxLines(config.maxLogLines);  // Keep last 100000 lines
   asyncFileLogger.startLogging();
   trantor::Logger::setOutputFunction(
       [&](const char* msg, const uint64_t len) {
-        asyncFileLogger.output(msg, len);
+        asyncFileLogger.output_(msg, len);
       },
       [&]() { asyncFileLogger.flush(); });
-  asyncFileLogger.setFileSizeLimit(cortex_utils::log_file_size_limit);
   // Number of cortex.cpp threads
   // if (argc > 1) {
   //   thread_num = std::atoi(argv[1]);
@@ -66,10 +69,10 @@ void RunServer() {
   LOG_INFO << "cortex.cpp version: undefined";
 #endif
 
-  LOG_INFO << "Server started, listening at: " << config.host << ":"
-           << config.port;
+  LOG_INFO << "Server started, listening at: " << config.apiServerHost << ":"
+           << config.apiServerPort;
   LOG_INFO << "Please load your model";
-  drogon::app().addListener(config.host, std::stoi(config.port));
+  drogon::app().addListener(config.apiServerHost, std::stoi(config.apiServerPort));
   drogon::app().setThreadNum(drogon_thread_num);
   LOG_INFO << "Number of thread is:" << drogon::app().getThreadNum();
 
@@ -154,6 +157,18 @@ int main(int argc, char* argv[]) {
       RunServer();
       return 0;
     } else {
+      auto config = file_manager_utils::GetCortexConfig();
+      trantor::FileLogger asyncFileLogger;
+      asyncFileLogger.setFileName(config.logFolderPath + "/" +
+                                  cortex_utils::logs_cli_base_name);
+      asyncFileLogger.setMaxLines(
+          config.maxLogLines);  // Keep last 100000 lines
+      asyncFileLogger.startLogging();
+      trantor::Logger::setOutputFunction(
+          [&](const char* msg, const uint64_t len) {
+            asyncFileLogger.output_(msg, len);
+          },
+          [&]() { asyncFileLogger.flush(); });
       CommandLineParser clp;
       clp.SetupCommand(argc, argv);
       return 0;
diff --git a/engine/services/download_service.cc b/engine/services/download_service.cc
index ffb25f02d..78a0c4757 100644
--- a/engine/services/download_service.cc
+++ b/engine/services/download_service.cc
@@ -81,6 +81,7 @@ void DownloadService::StartDownloadItem(
         }
         if (current == total) {
           outputFile.flush();
+          outputFile.close();
           CLI_LOG("Done download: " << static_cast<double>(total) / 1024 / 1024
                                     << " MiB");
           if (callback.has_value()) {
diff --git a/engine/test/components/test_cortex_upd_cmd.cc b/engine/test/components/test_cortex_upd_cmd.cc
new file mode 100644
index 000000000..9b8f3e764
--- /dev/null
+++ b/engine/test/components/test_cortex_upd_cmd.cc
@@ -0,0 +1,56 @@
+#include "commands/cortex_upd_cmd.h"
+#include "gtest/gtest.h"
+
+namespace {
+constexpr const auto kNewReleaseFolder = "./cortex-release";
+constexpr const auto kNewReleaseFile = "./cortex-release/cortexexe";
+constexpr const auto kCurReleaseFile = "./cortexexe";
+std::filesystem::path GetCortexTemp() {
+    return std::filesystem::temp_directory_path() / "cortex_temp";
+}
+}  // namespace
+
+class CortexUpdCmdTest : public ::testing::Test {
+  void SetUp() {
+    // Create new release folder and files
+    std::filesystem::path folder_path(kNewReleaseFolder);
+    std::filesystem::create_directory(folder_path);
+    std::ofstream src(kNewReleaseFile);
+    src.close();
+    std::ofstream dst(kCurReleaseFile);
+    dst.close();
+  }
+
+  void TearDown() {
+    std::filesystem::path folder_path(kNewReleaseFolder);
+    if (std::filesystem::exists(folder_path)) {
+      std::filesystem::remove_all(folder_path);
+    }
+
+    if (std::filesystem::exists(kCurReleaseFile)) {
+      std::filesystem::remove(kCurReleaseFile);
+    }
+
+    if (std::filesystem::exists(GetCortexTemp())) {
+      std::filesystem::remove(GetCortexTemp());
+    }
+  }
+};
+
+TEST_F(CortexUpdCmdTest, return_true_if_self_replace) {
+  EXPECT_TRUE(commands::ReplaceBinaryInflight("test", "test"));
+}
+
+TEST_F(CortexUpdCmdTest, replace_binary_successfully) {
+  std::filesystem::path new_binary(kNewReleaseFile);
+  std::filesystem::path cur_binary(kCurReleaseFile);
+  EXPECT_TRUE(commands::ReplaceBinaryInflight(new_binary, cur_binary));
+  EXPECT_TRUE(std::filesystem::exists(GetCortexTemp()));
+}
+
+TEST_F(CortexUpdCmdTest, should_restore_old_binary_if_has_error) {
+  std::filesystem::path new_binary("Non-exist");
+  std::filesystem::path cur_binary(kCurReleaseFile);
+  EXPECT_FALSE(commands::ReplaceBinaryInflight(new_binary, cur_binary));
+  EXPECT_FALSE(std::filesystem::exists(GetCortexTemp()));
+}
\ No newline at end of file
diff --git a/engine/test/components/test_engine_matcher_utils.cc b/engine/test/components/test_engine_matcher_utils.cc
new file mode 100644
index 000000000..d724c3fde
--- /dev/null
+++ b/engine/test/components/test_engine_matcher_utils.cc
@@ -0,0 +1,137 @@
+#include <string>
+#include <vector>
+#include "gtest/gtest.h"
+#include "utils/engine_matcher_utils.h"
+
+class EngineMatcherUtilsTestSuite : public ::testing::Test {
+ protected:
+  const std::vector<std::string> cortex_llamacpp_variants{
+      "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-avx-cuda-11-7.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-avx-cuda-12-0.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-avx.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-avx2-cuda-11-7.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-avx2-cuda-12-0.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-avx2.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-avx512-cuda-11-7.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-avx512-cuda-12-0.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-avx512.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-noavx-cuda-11-7.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-noavx-cuda-12-0.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-noavx.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-vulkan.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-mac-amd64.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-mac-arm64.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx-cuda-11-7.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx-cuda-12-0.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx2-cuda-11-7.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx2-cuda-12-0.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx2.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx512-cuda-11-7.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx512-cuda-12-0.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx512.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-noavx-cuda-11-7.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-noavx-cuda-12-0.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-noavx.tar.gz",
+      "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-vulkan.tar.gz",
+  };
+
+  const std::vector<std::string> cortex_tensorrt_variants{
+      "cortex.tensorrt-llm-0.0.9-linux-cuda-12-4.tar.gz",
+      "cortex.tensorrt-llm-0.0.9-windows-cuda-12-4.tar.gz"};
+
+  const std::vector<std::string> cortex_onnx_variants{
+      "cortex.onnx-0.1.7-windows-amd64.tar.gz"};
+};
+
+TEST_F(EngineMatcherUtilsTestSuite, TestValidateOnnx) {
+
+  {
+    auto expect_matched_variant = cortex_onnx_variants[0];
+    auto result = engine_matcher_utils::ValidateOnnx(cortex_onnx_variants,
+                                                     "windows", "amd64");
+
+    EXPECT_EQ(result, expect_matched_variant);
+  }
+
+  {
+    // should return an empty variant because no variant matched
+    auto expect_matched_variant{""};
+    auto windows_arm_result = engine_matcher_utils::ValidateOnnx(
+        cortex_onnx_variants, "windows", "arm");
+    auto mac_arm64_result = engine_matcher_utils::ValidateOnnx(
+        cortex_onnx_variants, "mac", "arm64");
+
+    EXPECT_EQ(windows_arm_result, expect_matched_variant);
+    EXPECT_EQ(mac_arm64_result, expect_matched_variant);
+  }
+}
+
+TEST_F(EngineMatcherUtilsTestSuite, TestValidateTensorrt) {
+
+  {
+    auto windows_expect_matched_variant{cortex_tensorrt_variants[1]};
+    auto linux_expect_matched_variant{cortex_tensorrt_variants[0]};
+    auto windows{"windows"};
+    auto linux{"linux"};
+    auto cuda_version{"12.4"};
+    auto windows_result = engine_matcher_utils::ValidateTensorrtLlm(
+        cortex_tensorrt_variants, windows, cuda_version);
+    auto linux_result = engine_matcher_utils::ValidateTensorrtLlm(
+        cortex_tensorrt_variants, linux, cuda_version);
+
+    EXPECT_EQ(windows_result, windows_expect_matched_variant);
+    EXPECT_EQ(linux_result, linux_expect_matched_variant);
+  }
+
+  {  // macos is not supported
+    auto os = "mac";
+    auto cuda_version{"12.4"};
+
+    auto result = engine_matcher_utils::ValidateTensorrtLlm(
+        cortex_tensorrt_variants, os, cuda_version);
+    EXPECT_EQ(result, "");
+  }
+}
+
+TEST_F(EngineMatcherUtilsTestSuite, TestValidate) {
+  {
+    auto os{"windows"};
+    auto cpu_arch{"amd64"};
+    auto suitable_avx{"avx2"};
+    auto cuda_version{"12.4"};
+
+    auto variant = engine_matcher_utils::Validate(
+        cortex_llamacpp_variants, os, cpu_arch, suitable_avx, cuda_version);
+
+    EXPECT_EQ(
+        variant,
+        "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx2-cuda-12-0.tar.gz");
+  }
+
+  {
+    auto os{"mac"};
+    auto cpu_arch{"amd64"};
+    auto suitable_avx{""};
+    auto cuda_version{""};
+
+    auto variant = engine_matcher_utils::Validate(
+        cortex_llamacpp_variants, os, cpu_arch, suitable_avx, cuda_version);
+
+    EXPECT_EQ(variant, "cortex.llamacpp-0.1.25-25.08.24-mac-amd64.tar.gz");
+  }
+
+  {
+    auto os{"windows"};
+    auto cpu_arch{"amd64"};
+    auto suitable_avx{"avx2"};
+    auto cuda_version{"10"};
+
+    auto variant = engine_matcher_utils::Validate(
+        cortex_llamacpp_variants, os, cpu_arch, suitable_avx, cuda_version);
+
+    // fallback to no cuda version
+    EXPECT_EQ(variant,
+              "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx2.tar.gz");
+  }
+}
diff --git a/engine/test/components/test_system_info_utils.cc b/engine/test/components/test_system_info_utils.cc
new file mode 100644
index 000000000..da38bde5d
--- /dev/null
+++ b/engine/test/components/test_system_info_utils.cc
@@ -0,0 +1,90 @@
+#include <string>
+#include "gtest/gtest.h"
+
+class SystemInfoUtilsTestSuite : public ::testing::Test {
+ protected:
+  const std::string nvidia_smi_sample_output = R"(
+Sun Aug 25 22:29:25 2024
++-----------------------------------------------------------------------------------------+
+| NVIDIA-SMI 560.70                 Driver Version: 560.70         CUDA Version: 12.6     |
+|-----------------------------------------+------------------------+----------------------+
+| GPU  Name                  Driver-Model | Bus-Id          Disp.A | Volatile Uncorr. ECC |
+| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
+|                                         |                        |               MIG M. |
+|=========================================+========================+======================|
+|   0  NVIDIA GeForce RTX 4090      WDDM  |   00000000:01:00.0 Off |                  Off |
+|  0%   24C    P8             10W /  500W |     395MiB /  24564MiB |     19%      Default |
+|                                         |                        |                  N/A |
++-----------------------------------------+------------------------+----------------------+
+
++-----------------------------------------------------------------------------------------+
+| Processes:                                                                              |
+|  GPU   GI   CI        PID   Type   Process name                              GPU Memory |
+|        ID   ID                                                               Usage      |
+|=========================================================================================|
+|    0   N/A  N/A      3984    C+G   ...5n1h2txyewy\ShellExperienceHost.exe      N/A      |
+|    0   N/A  N/A      7904    C+G   ...ekyb3d8bbwe\PhoneExperienceHost.exe      N/A      |
+|    0   N/A  N/A      8240    C+G   ...__8wekyb3d8bbwe\WindowsTerminal.exe      N/A      |
+|    0   N/A  N/A      8904    C+G   C:\Windows\explorer.exe                     N/A      |
+|    0   N/A  N/A      9304    C+G   ...siveControlPanel\SystemSettings.exe      N/A      |
+|    0   N/A  N/A      9944    C+G   ...nt.CBS_cw5n1h2txyewy\SearchHost.exe      N/A      |
+|    0   N/A  N/A     11140    C+G   ...2txyewy\StartMenuExperienceHost.exe      N/A      |
++-----------------------------------------------------------------------------------------+
+)";
+
+  const std::string vulkan_sample_output = R"(
+==========
+VULKANINFO
+==========
+
+Vulkan Instance Version: 1.3.280
+
+
+Instance Extensions: count = 19
+-------------------------------
+VK_EXT_debug_report                    : extension revision 10
+VK_EXT_debug_utils                     : extension revision 2
+VK_EXT_direct_mode_display             : extension revision 1
+VK_EXT_surface_maintenance1            : extension revision 1
+VK_EXT_swapchain_colorspace            : extension revision 4
+VK_KHR_device_group_creation           : extension revision 1
+VK_KHR_display                         : extension revision 23
+VK_KHR_external_fence_capabilities     : extension revision 1
+VK_KHR_external_memory_capabilities    : extension revision 1
+VK_KHR_external_semaphore_capabilities : extension revision 1
+VK_KHR_get_display_properties2         : extension revision 1
+VK_KHR_get_physical_device_properties2 : extension revision 2
+VK_KHR_get_surface_capabilities2       : extension revision 1
+VK_KHR_portability_enumeration         : extension revision 1
+VK_KHR_surface                         : extension revision 25
+VK_KHR_surface_protected_capabilities  : extension revision 1
+VK_KHR_win32_surface                   : extension revision 6
+VK_LUNARG_direct_driver_loading        : extension revision 1
+VK_NV_external_memory_capabilities     : extension revision 1
+
+Instance Layers: count = 1
+--------------------------
+VK_LAYER_NV_optimus NVIDIA Optimus layer 1.3.280  version 1
+
+Devices:
+========
+GPU0:
+        apiVersion         = 1.3.280
+        driverVersion      = 560.70.0.0
+        vendorID           = 0x10de
+        deviceID           = 0x2684
+        deviceType         = PHYSICAL_DEVICE_TYPE_DISCRETE_GPU
+        deviceName         = NVIDIA GeForce RTX 4090
+        driverID           = DRIVER_ID_NVIDIA_PROPRIETARY
+        driverName         = NVIDIA
+        driverInfo         = 560.70
+        conformanceVersion = 1.3.8.2
+        deviceUUID         = 11deafdf-9f15-e857-2a87-8acc153fc9f7
+        driverUUID         = 10f251d9-d3c0-5001-bf67-24bb06423040
+)";
+
+  const std::string gpu_query_list_sample_output = R"(
+0, 46068, NVIDIA RTX A6000, 8.6
+1, 46068, NVIDIA RTX A6000, 8.6
+)";
+};
diff --git a/engine/utils/config_yaml_utils.h b/engine/utils/config_yaml_utils.h
index 4330f3527..8e3668292 100644
--- a/engine/utils/config_yaml_utils.h
+++ b/engine/utils/config_yaml_utils.h
@@ -8,14 +8,17 @@
 
 namespace config_yaml_utils {
 struct CortexConfig {
+  std::string logFolderPath;
   std::string dataFolderPath;
-  std::string host;
-  std::string port;
+  int maxLogLines;
+  std::string apiServerHost;
+  std::string apiServerPort;
 };
 
 const std::string kCortexFolderName = "cortexcpp";
 const std::string kDefaultHost{"127.0.0.1"};
 const std::string kDefaultPort{"3928"};
+const int kDefaultMaxLines{100000};
 
 inline void DumpYamlConfig(const CortexConfig& config,
                            const std::string& path) {
@@ -27,9 +30,11 @@ inline void DumpYamlConfig(const CortexConfig& config,
       throw std::runtime_error("Failed to open output file.");
     }
     YAML::Node node;
+    node["logFolderPath"] = config.logFolderPath;
     node["dataFolderPath"] = config.dataFolderPath;
-    node["host"] = config.host;
-    node["port"] = config.port;
+    node["maxLogLines"] = config.maxLogLines;
+    node["apiServerHost"] = config.apiServerHost;
+    node["apiServerPort"] = config.apiServerPort;
 
     out_file << node;
     out_file.close();
@@ -48,10 +53,18 @@ inline CortexConfig FromYaml(const std::string& path,
 
   try {
     auto node = YAML::LoadFile(config_file_path.string());
+    int max_lines;
+    if (!node["maxLogLines"]) {
+      max_lines = kDefaultMaxLines;
+    } else {
+      max_lines = node["maxLogLines"].as<int>();
+    }
     CortexConfig config = {
+        .logFolderPath = node["logFolderPath"].as<std::string>(),
         .dataFolderPath = node["dataFolderPath"].as<std::string>(),
-        .host = node["host"].as<std::string>(),
-        .port = node["port"].as<std::string>(),
+        .maxLogLines = max_lines,
+        .apiServerHost = node["apiServerHost"].as<std::string>(),
+        .apiServerPort = node["apiServerPort"].as<std::string>(),
     };
     return config;
   } catch (const YAML::BadFile& e) {
@@ -60,5 +73,4 @@ inline CortexConfig FromYaml(const std::string& path,
   }
 }
 
-
 }  // namespace config_yaml_utils
diff --git a/engine/utils/cortex_utils.h b/engine/utils/cortex_utils.h
index 32dea9321..9673f0c1a 100644
--- a/engine/utils/cortex_utils.h
+++ b/engine/utils/cortex_utils.h
@@ -34,8 +34,8 @@ constexpr static auto kTensorrtLlmPath = "/engines/cortex.tensorrt-llm";
 
 inline std::string models_folder = "./models";
 inline std::string logs_folder = "./logs";
-inline std::string logs_base_name = "./logs/cortex";
-inline size_t log_file_size_limit = 20000000;  // ~20 mb
+inline std::string logs_base_name = "./logs/cortex.log";
+inline std::string logs_cli_base_name = "./logs/cortex-cli.log";
 
 inline std::string extractBase64(const std::string& input) {
   std::regex pattern("base64,(.*)");
diff --git a/engine/utils/engine_matcher_utils.h b/engine/utils/engine_matcher_utils.h
index 77baf1f72..287304f02 100644
--- a/engine/utils/engine_matcher_utils.h
+++ b/engine/utils/engine_matcher_utils.h
@@ -8,49 +8,10 @@
 #include "utils/logging_utils.h"
 
 namespace engine_matcher_utils {
-// for testing purpose
-const std::vector<std::string> cortex_llamacpp_variants{
-    "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-avx-cuda-11-7.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-avx-cuda-12-0.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-avx.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-avx2-cuda-11-7.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-avx2-cuda-12-0.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-avx2.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-avx512-cuda-11-7.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-avx512-cuda-12-0.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-avx512.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-noavx-cuda-11-7.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-noavx-cuda-12-0.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-noavx.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-linux-amd64-vulkan.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-mac-amd64.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-mac-arm64.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx-cuda-11-7.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx-cuda-12-0.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx2-cuda-11-7.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx2-cuda-12-0.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx2.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx512-cuda-11-7.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx512-cuda-12-0.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-avx512.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-noavx-cuda-11-7.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-noavx-cuda-12-0.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-noavx.tar.gz",
-    "cortex.llamacpp-0.1.25-25.08.24-windows-amd64-vulkan.tar.gz",
-};
-const std::vector<std::string> cortex_onnx_variants{
-    "cortex.onnx-0.1.7-windows-amd64.tar.gz"};
-
-const std::vector<std::string> cortex_tensorrt_variants{
-    "cortex.tensorrt-llm-0.0.9-linux-cuda-12-4.tar.gz",
-    "cortex.tensorrt-llm-0.0.9-windows-cuda-12-4.tar.gz"};
-
-inline std::string GetSuitableAvxVariant() {
-  cortex::cpuid::CpuInfo cpu_info;
-
+inline std::string GetSuitableAvxVariant(cortex::cpuid::CpuInfo& cpu_info) {
   CTL_INF("GetSuitableAvxVariant:" << "\n" << cpu_info.to_string());
 
+  // prioritize avx512 > avx2 > avx > noavx
   if (cpu_info.has_avx512_f())
     return "avx512";
   if (cpu_info.has_avx2())
@@ -151,10 +112,8 @@ inline std::string Validate(const std::vector<std::string>& variants,
                             const std::string& os, const std::string& cpu_arch,
                             const std::string& suitable_avx,
                             const std::string& cuda_version) {
-
-  // Early return if the OS is unsupported
+  // Early return if the OS is not supported
   if (os != "mac" && os != "windows" && os != "linux") {
-    // TODO: throw is better
     return "";
   }
 
diff --git a/engine/utils/file_logger.cc b/engine/utils/file_logger.cc
new file mode 100644
index 000000000..f684d457f
--- /dev/null
+++ b/engine/utils/file_logger.cc
@@ -0,0 +1,175 @@
+#include "file_logger.h"
+#include <algorithm>
+#include <iostream>
+#include <sstream>
+
+#ifdef _WIN32
+#include <io.h>
+#define ftruncate _chsize
+#else
+#include <unistd.h>
+#endif
+#include <string.h>
+
+using namespace trantor;
+
+FileLogger::FileLogger() : AsyncFileLogger() {}
+
+FileLogger::~FileLogger() = default;
+
+void FileLogger::output_(const char* msg, const uint64_t len) {
+  if (!circular_log_file_ptr_) {
+    circular_log_file_ptr_ =
+        std::make_unique<CircularLogFile>(fileBaseName_, max_lines_);
+  }
+  circular_log_file_ptr_->writeLog(msg, len);
+}
+
+FileLogger::CircularLogFile::CircularLogFile(const std::string& fileName,
+                                             uint64_t maxLines)
+    : max_lines_(maxLines), file_name_(fileName) {
+  std::lock_guard<std::mutex> lock(mutex_);
+  OpenFile();
+  LoadExistingLines();
+  TruncateFileIfNeeded();
+}
+
+FileLogger::CircularLogFile::~CircularLogFile() {
+  std::lock_guard<std::mutex> lock(mutex_);
+  CloseFile();
+}
+void FileLogger::CircularLogFile::writeLog(const char* logLine,
+                                           const uint64_t len) {
+  std::lock_guard<std::mutex> lock(mutex_);
+  if (!fp_)
+    return;
+
+  std::string logString(logLine, len);
+  std::istringstream iss(logString);
+  std::string line;
+  while (std::getline(iss, line)) {
+    if (lineBuffer_.size() >= max_lines_) {
+      lineBuffer_.pop_front();
+    }
+    lineBuffer_.push_back(line);
+    AppendToFile(line + "\n");
+    ++linesWrittenSinceLastTruncate_;
+    if (linesWrittenSinceLastTruncate_.load() >= TRUNCATE_CHECK_INTERVAL) {
+
+      TruncateFileIfNeeded();
+    }
+  }
+}
+void FileLogger::CircularLogFile::flush() {
+  std::lock_guard<std::mutex> lock(mutex_);
+  if (fp_) {
+    fflush(fp_);
+  }
+}
+
+void FileLogger::CircularLogFile::TruncateFileIfNeeded() {
+  // std::cout<<"Truncating file "<< totalLines_ <<std::endl;
+  if (!fp_ || lineBuffer_.size() < max_lines_)
+    return;
+
+  // Close the current file
+  fclose(fp_);
+  fp_ = nullptr;
+
+  // Open a temporary file for writing
+  std::string tempFileName = file_name_ + ".temp";
+  FILE* tempFile = fopen(tempFileName.c_str(), "w");
+  if (!tempFile) {
+
+    std::cout << "Error opening temporary file for truncation: "
+              << strerror(errno) << std::endl;
+    OpenFile();  // Reopen the original file
+    return;
+  }
+
+  // Write only the last max_lines_ lines to the temporary file
+  size_t startIndex =
+      lineBuffer_.size() > max_lines_ ? lineBuffer_.size() - max_lines_ : 0;
+
+  for (size_t i = startIndex; i < lineBuffer_.size(); ++i) {
+    fprintf(tempFile, "%s\n", lineBuffer_[i].c_str());
+  }
+
+  fclose(tempFile);
+
+  // Replace the original file with the temporary file
+  if (std::rename(tempFileName.c_str(), file_name_.c_str()) != 0) {
+    std::cout << "Error replacing original file with truncated file: "
+              << strerror(errno) << std::endl;
+    std::remove(tempFileName.c_str());  // Clean up the temporary file
+  }
+  //    else {
+  //     totalLines_.store(lineBuffer_.size() > max_lines_ ? max_lines_
+  //                                                       : lineBuffer_.size());
+  //   }
+
+  // Reopen the file
+  OpenFile();
+  // LoadExistingLines();
+  linesWrittenSinceLastTruncate_.store(0);
+}
+
+void FileLogger::CircularLogFile::OpenFile() {
+#ifdef _WIN32
+  auto wFileName = utils::toNativePath(file_name_);
+  fp_ = _wfopen(wFileName.c_str(), L"r+");
+#else
+  fp_ = fopen(file_name_.c_str(), "r+");
+#endif
+
+  if (!fp_) {
+// If file doesn't exist, create it
+#ifdef _WIN32
+    fp_ = _wfopen(wFileName.c_str(), L"w+");
+#else
+    fp_ = fopen(file_name_.c_str(), "w+");
+#endif
+
+    if (!fp_) {
+      std::cerr << "Error opening file: " << strerror(errno) << std::endl;
+    }
+  }
+}
+void FileLogger::CircularLogFile::LoadExistingLines() {
+  if (!fp_)
+    return;
+
+  // Move to the beginning of the file
+  fseek(fp_, 0, SEEK_SET);
+
+  lineBuffer_.clear();
+
+  std::string line;
+  char buffer[4096];
+  while (fgets(buffer, sizeof(buffer), fp_) != nullptr) {
+    line = buffer;
+    if (!line.empty() && line.back() == '\n') {
+      line.pop_back();  // Remove trailing newline
+    }
+    if (lineBuffer_.size() >= max_lines_) {
+      lineBuffer_.pop_front();
+    }
+    lineBuffer_.push_back(line);
+  }
+
+  // Move back to the end of the file for appending
+  fseek(fp_, 0, SEEK_END);
+}
+void FileLogger::CircularLogFile::AppendToFile(const std::string& line) {
+  if (fp_) {
+    fwrite(line.c_str(), 1, line.length(), fp_);
+    fflush(fp_);
+  }
+}
+
+void FileLogger::CircularLogFile::CloseFile() {
+  if (fp_) {
+    fclose(fp_);
+    fp_ = nullptr;
+  }
+}
\ No newline at end of file
diff --git a/engine/utils/file_logger.h b/engine/utils/file_logger.h
new file mode 100644
index 000000000..58b719019
--- /dev/null
+++ b/engine/utils/file_logger.h
@@ -0,0 +1,72 @@
+#pragma once
+
+#include <trantor/utils/AsyncFileLogger.h>
+#include <trantor/utils/Utilities.h>
+#include <atomic>
+#include <cstddef>
+#include <cstdint>
+#include <deque>
+#include <mutex>
+
+#ifdef _WIN32
+#include <windows.h>
+#else
+#include <unistd.h>
+#endif
+
+namespace trantor {
+
+class TRANTOR_EXPORT FileLogger : public AsyncFileLogger {
+ public:
+  FileLogger();
+  ~FileLogger();
+
+  /**
+     * @brief Set the maximum number of lines to keep in the log file.
+     *
+     * @param maxLines
+     */
+  void setMaxLines(uint64_t maxLines) { max_lines_ = maxLines; }
+
+  /**
+     * @brief Set the log file name.
+     *
+     * @param fileName The full name of the log file.
+     */
+  void setFileName(const std::string& fileName) {
+    filePath_ = "./";
+    fileBaseName_ = fileName;
+    fileExtName_ = "";
+  }
+  void output_(const char* msg, const uint64_t len);
+
+ protected:
+  class CircularLogFile {
+   public:
+    CircularLogFile(const std::string& fileName, uint64_t maxLines);
+    ~CircularLogFile();
+
+    void writeLog(const char* logLine, const uint64_t len);
+    void flush();
+    uint64_t getLength() const { return lineBuffer_.size(); }
+
+   private:
+    FILE* fp_{nullptr};
+    uint64_t max_lines_;
+    std::string file_name_;
+    std::deque<std::string> lineBuffer_;
+    std::atomic<int> linesWrittenSinceLastTruncate_{0};
+    static const uint64_t TRUNCATE_CHECK_INTERVAL = 1000;
+    mutable std::mutex mutex_;
+
+    void LoadExistingLines();
+    void TruncateFileIfNeeded();
+    void AppendToFile(const std::string& line);
+    void OpenFile();
+    void CloseFile();
+  };
+  std::unique_ptr<CircularLogFile> circular_log_file_ptr_;
+  uint64_t max_lines_{100000};  // Default to 100000 lines
+};
+
+}  // namespace trantor
\ No newline at end of file
diff --git a/engine/utils/file_manager_utils.h b/engine/utils/file_manager_utils.h
index d20a39f13..b4d7ab07a 100644
--- a/engine/utils/file_manager_utils.h
+++ b/engine/utils/file_manager_utils.h
@@ -135,9 +135,11 @@ inline void CreateConfigFileIfNotExist() {
   CTL_INF("Default data folder path: " + defaultDataFolderPath.string());
 
   auto config = config_yaml_utils::CortexConfig{
+      .logFolderPath = defaultDataFolderPath.string(),
       .dataFolderPath = defaultDataFolderPath.string(),
-      .host = config_yaml_utils::kDefaultHost,
-      .port = config_yaml_utils::kDefaultPort,
+      .maxLogLines = config_yaml_utils::kDefaultMaxLines,
+      .apiServerHost = config_yaml_utils::kDefaultHost,
+      .apiServerPort = config_yaml_utils::kDefaultPort,
   };
   DumpYamlConfig(config, config_path.string());
 }
@@ -169,6 +171,27 @@ inline std::filesystem::path GetCortexDataPath() {
   return data_folder_path;
 }
 
+inline std::filesystem::path GetCortexLogPath() {
+  // TODO: We will need to support user to move the data folder to other place.
+  // TODO: get the variant of cortex. As discussed, we will have: prod, beta, nightly
+  // currently we will store cortex data at ~/cortexcpp
+  auto config = GetCortexConfig();
+  std::filesystem::path log_folder_path;
+  if (!config.logFolderPath.empty()) {
+    log_folder_path = std::filesystem::path(config.logFolderPath);
+  } else {
+    auto home_path = GetHomeDirectoryPath();
+    log_folder_path = home_path / config_yaml_utils::kCortexFolderName;
+  }
+
+  if (!std::filesystem::exists(log_folder_path)) {
+    CTL_INF("Cortex log folder not found. Create one: " +
+            log_folder_path.string());
+    std::filesystem::create_directory(log_folder_path);
+  }
+  return log_folder_path;
+}
+
 inline std::filesystem::path GetModelsContainerPath() {
   auto cortex_path = GetCortexDataPath();
   auto models_container_path = cortex_path / "models";
diff --git a/engine/utils/http_util.h b/engine/utils/http_util.h
index 471ef3b27..422d2c543 100644
--- a/engine/utils/http_util.h
+++ b/engine/utils/http_util.h
@@ -1,6 +1,7 @@
 #pragma once
 
 #include <drogon/HttpController.h>
+#include "utils/cortex_utils.h"
 
 using namespace drogon;
 
@@ -21,4 +22,4 @@ inline bool HasFieldInReq(const HttpRequestPtr& req,
   return true;
 }
 
-}  // namespace http_util
\ No newline at end of file
+}  // namespace http_util
diff --git a/engine/utils/logging_utils.h b/engine/utils/logging_utils.h
index 77311c13e..fcaa3f4bb 100644
--- a/engine/utils/logging_utils.h
+++ b/engine/utils/logging_utils.h
@@ -30,4 +30,3 @@ inline bool log_verbose = false;
   } else {                         \
     std::cout << msg << std::endl; \
   }
-
diff --git a/engine/utils/system_info_utils.h b/engine/utils/system_info_utils.h
index ede7915d2..9cdcc8f05 100644
--- a/engine/utils/system_info_utils.h
+++ b/engine/utils/system_info_utils.h
@@ -79,91 +79,6 @@ inline SystemInfo GetSystemInfo() {
   return SystemInfo{os.str(), arch.str()};
 }
 
-constexpr auto vulkan_sample_output = R"(
-==========
-VULKANINFO
-==========
-
-Vulkan Instance Version: 1.3.280
-
-
-Instance Extensions: count = 19
--------------------------------
-VK_EXT_debug_report                    : extension revision 10
-VK_EXT_debug_utils                     : extension revision 2
-VK_EXT_direct_mode_display             : extension revision 1
-VK_EXT_surface_maintenance1            : extension revision 1
-VK_EXT_swapchain_colorspace            : extension revision 4
-VK_KHR_device_group_creation           : extension revision 1
-VK_KHR_display                         : extension revision 23
-VK_KHR_external_fence_capabilities     : extension revision 1
-VK_KHR_external_memory_capabilities    : extension revision 1
-VK_KHR_external_semaphore_capabilities : extension revision 1
-VK_KHR_get_display_properties2         : extension revision 1
-VK_KHR_get_physical_device_properties2 : extension revision 2
-VK_KHR_get_surface_capabilities2       : extension revision 1
-VK_KHR_portability_enumeration         : extension revision 1
-VK_KHR_surface                         : extension revision 25
-VK_KHR_surface_protected_capabilities  : extension revision 1
-VK_KHR_win32_surface                   : extension revision 6
-VK_LUNARG_direct_driver_loading        : extension revision 1
-VK_NV_external_memory_capabilities     : extension revision 1
-
-Instance Layers: count = 1
---------------------------
-VK_LAYER_NV_optimus NVIDIA Optimus layer 1.3.280  version 1
-
-Devices:
-========
-GPU0:
-        apiVersion         = 1.3.280
-        driverVersion      = 560.70.0.0
-        vendorID           = 0x10de
-        deviceID           = 0x2684
-        deviceType         = PHYSICAL_DEVICE_TYPE_DISCRETE_GPU
-        deviceName         = NVIDIA GeForce RTX 4090
-        driverID           = DRIVER_ID_NVIDIA_PROPRIETARY
-        driverName         = NVIDIA
-        driverInfo         = 560.70
-        conformanceVersion = 1.3.8.2
-        deviceUUID         = 11deafdf-9f15-e857-2a87-8acc153fc9f7
-        driverUUID         = 10f251d9-d3c0-5001-bf67-24bb06423040
-)";
-
-constexpr auto gpu_query_list_sample_output = R"(
-0, 46068, NVIDIA RTX A6000, 8.6
-1, 46068, NVIDIA RTX A6000, 8.6
-)";
-
-constexpr auto nvidia_smi_sample_output = R"(
-Sun Aug 25 22:29:25 2024
-+-----------------------------------------------------------------------------------------+
-| NVIDIA-SMI 560.70                 Driver Version: 560.70         CUDA Version: 12.6     |
-|-----------------------------------------+------------------------+----------------------+
-| GPU  Name                  Driver-Model | Bus-Id          Disp.A | Volatile Uncorr. ECC |
-| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
-|                                         |                        |               MIG M. |
-|=========================================+========================+======================|
-|   0  NVIDIA GeForce RTX 4090      WDDM  |   00000000:01:00.0 Off |                  Off |
-|  0%   24C    P8             10W /  500W |     395MiB /  24564MiB |     19%      Default |
-|                                         |                        |                  N/A |
-+-----------------------------------------+------------------------+----------------------+
-
-+-----------------------------------------------------------------------------------------+
-| Processes:                                                                              |
-|  GPU   GI   CI        PID   Type   Process name                              GPU Memory |
-|        ID   ID                                                               Usage      |
-|=========================================================================================|
-|    0   N/A  N/A      3984    C+G   ...5n1h2txyewy\ShellExperienceHost.exe      N/A      |
-|    0   N/A  N/A      7904    C+G   ...ekyb3d8bbwe\PhoneExperienceHost.exe      N/A      |
-|    0   N/A  N/A      8240    C+G   ...__8wekyb3d8bbwe\WindowsTerminal.exe      N/A      |
-|    0   N/A  N/A      8904    C+G   C:\Windows\explorer.exe                     N/A      |
-|    0   N/A  N/A      9304    C+G   ...siveControlPanel\SystemSettings.exe      N/A      |
-|    0   N/A  N/A      9944    C+G   ...nt.CBS_cw5n1h2txyewy\SearchHost.exe      N/A      |
-|    0   N/A  N/A     11140    C+G   ...2txyewy\StartMenuExperienceHost.exe      N/A      |
-+-----------------------------------------------------------------------------------------+
-)";
-
 inline bool IsNvidiaSmiAvailable() {
 #ifdef _WIN32
   // Check if nvidia-smi.exe exists in the PATH on Windows
diff --git a/engine/vcpkg.json b/engine/vcpkg.json
index 74f89965a..fe4783ec8 100644
--- a/engine/vcpkg.json
+++ b/engine/vcpkg.json
@@ -1,5 +1,6 @@
 {
     "dependencies": [
+      "gtest",
       "cli11",
       {
         "name": "cpp-httplib",
@@ -16,4 +17,4 @@
       "libarchive",
       "tabulate"
     ]
-  }
\ No newline at end of file
+}