diff --git a/.github/workflows/build-nitro-node.yml b/.github/workflows/build-nitro-node.yml
index 4fe80b692..7ebdb6fde 100644
--- a/.github/workflows/build-nitro-node.yml
+++ b/.github/workflows/build-nitro-node.yml
@@ -13,189 +13,24 @@ on:
     paths: [".github/workflows/build-nitro-node.yml", "nitro-node"]
   workflow_dispatch:
 
-jobs:
-  ubuntu-amd64-non-cuda-build:
-    runs-on: ubuntu-latest
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-        with:
-          submodules: recursive
+env:
+  LLM_MODEL_URL: https://huggingface.co/TheBloke/TinyLlama-1.1B-Chat-v0.3-GGUF/resolve/main/tinyllama-1.1b-chat-v0.3.Q2_K.gguf
+  WHISPER_MODEL_URL: https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-tiny-q5_1.bin
 
-      - uses: actions/setup-node@v4
-        with:
-          node-version: 18
-
-      - name: Restore cached model file
-        id: cache-model-restore
-        uses: actions/cache/restore@v4
-        with:
-          path: |
-            nitro-node/test/test_assets/*.gguf
-          key: ${{ runner.os }}-model-gguf
-
-      - uses: suisei-cn/actions-download-file@v1.4.0
-        id: download-model-file
-        name: Download model file
-        with:
-          url: "The model we are using is [tinyllama-1.1b](https://huggingface.co/TheBloke/TinyLlama-1.1B-Chat-v1.0-GGUF/resolve/main/tinyllama-1.1b-chat-v1.0.Q4_K_M.gguf)!"
-          target: nitro-node/test/test_assets/
-          auto-match: true
-          retry-times: 3
-
-      - name: Save downloaded model file to cache
-        id: cache-model-save
-        uses: actions/cache/save@v4
-        with:
-          path: |
-            nitro-node/test/test_assets/*.gguf
-          key: ${{ steps.cache-model-restore.outputs.cache-primary-key }}
-
-      - name: Run tests
-        id: test_nitro_node
-        run: |
-          cd nitro-node
-          make clean test-ci
-
-  #ubuntu-amd64-build:
-  #  runs-on: ubuntu-18-04-cuda-11-7
-  #  steps:
-  #    - name: Clone
-  #      id: checkout
-  #      uses: actions/checkout@v4
-  #      with:
-  #        submodules: recursive
-
-  #    - uses: actions/setup-node@v4
-  #      with:
-  #        node-version: 18
-
-  #    - name: Restore cached model file
-  #      id: cache-model-restore
-  #      uses: actions/cache/restore@v4
-  #      with:
-  #        path: |
-  #          nitro-node/test/test_assets/*.gguf
-  #        key: ${{ runner.os }}-model-gguf
-
-  #    - uses: suisei-cn/actions-download-file@v1.4.0
-  #      id: download-model-file
-  #      name: Download model file
-  #      with:
-  #        url: "The model we are using is [tinyllama-1.1b](https://huggingface.co/TheBloke/TinyLlama-1.1B-Chat-v1.0-GGUF/resolve/main/tinyllama-1.1b-chat-v1.0.Q4_K_M.gguf)!"
-  #        target: nitro-node/test/test_assets/
-  #        auto-match: true
-  #        retry-times: 3
-
-  #    - name: Save downloaded model file to cache
-  #      id: cache-model-save
-  #      uses: actions/cache/save@v4
-  #      with:
-  #        path: |
-  #          nitro-node/test/test_assets/*.gguf
-  #        key: ${{ steps.cache-model-restore.outputs.cache-primary-key }}
-
-  #    - name: Run tests
-  #      id: test_nitro_node
-  #      run: |
-  #        cd nitro-node
-  #        make clean test-ci
-
-  #ubuntu-amd64-cuda-build:
-  #  runs-on: ubuntu-18-04-cuda-${{ matrix.cuda }}
-  #  strategy:
-  #    matrix:
-  #      cuda: ["12-0", "11-7"]
-
-  #  steps:
-  #    - name: Clone
-  #      id: checkout
-  #      uses: actions/checkout@v4
-  #      with:
-  #        submodules: recursive
-
-  #    - uses: actions/setup-node@v4
-  #      with:
-  #        node-version: 18
-
-  #    - name: Restore cached model file
-  #      id: cache-model-restore
-  #      uses: actions/cache/restore@v4
-  #      with:
-  #        path: |
-  #          nitro-node/test/test_assets/*.gguf
-  #        key: ${{ runner.os }}-model-gguf
-
-  #    - uses: suisei-cn/actions-download-file@v1.4.0
-  #      id: download-model-file
-  #      name: Download model file
-  #      with:
-  #        url: "The model we are using is [tinyllama-1.1b](https://huggingface.co/TheBloke/TinyLlama-1.1B-Chat-v1.0-GGUF/resolve/main/tinyllama-1.1b-chat-v1.0.Q4_K_M.gguf)!"
-  #        target: nitro-node/test/test_assets/
-  #        auto-match: true
-  #        retry-times: 3
-
-  #    - name: Save downloaded model file to cache
-  #      id: cache-model-save
-  #      uses: actions/cache/save@v4
-  #      with:
-  #        path: |
-  #          nitro-node/test/test_assets/*.gguf
-  #        key: ${{ steps.cache-model-restore.outputs.cache-primary-key }}
-
-  #    - name: Run tests
-  #      id: test_nitro_node
-  #      run: |
-  #        cd nitro-node
-  #        make clean test-ci
-
-  macOS-M-build:
-    runs-on: macos-14
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-        with:
-          submodules: recursive
-
-      - uses: actions/setup-node@v4
-        with:
-          node-version: 18
-
-      - name: Restore cached model file
-        id: cache-model-restore
-        uses: actions/cache/restore@v4
-        with:
-          path: |
-            nitro-node/test/test_assets/*.gguf
-          key: ${{ runner.os }}-model-gguf
-
-      - uses: suisei-cn/actions-download-file@v1.4.0
-        id: download-model-file
-        name: Download model file
-        with:
-          url: "The model we are using is [tinyllama-1.1b](https://huggingface.co/TheBloke/TinyLlama-1.1B-Chat-v1.0-GGUF/resolve/main/tinyllama-1.1b-chat-v1.0.Q4_K_M.gguf)!"
-          target: nitro-node/test/test_assets/
-          auto-match: true
-          retry-times: 3
-
-      - name: Save downloaded model file to cache
-        id: cache-model-save
-        uses: actions/cache/save@v4
-        with:
-          path: |
-            nitro-node/test/test_assets/*.gguf
-          key: ${{ steps.cache-model-restore.outputs.cache-primary-key }}
-
-      - name: Run tests
-        id: test_nitro_node
-        run: |
-          cd nitro-node
-          make clean test-ci
+jobs:
+  build:
+    runs-on: ${{ matrix.os }}
+    strategy:
+      matrix:
+        os:
+          - ubuntu-18-04-cuda-12-0
+          - ubuntu-18-04-cuda-11-7
+          - macos-silicon
+          - macos-latest
+          - windows-latest
+          - windows-cuda-12-0
+          - windows-cuda-11-7
 
-  macOS-Intel-build:
-    runs-on: macos-latest
     steps:
       - name: Clone
         id: checkout
@@ -219,7 +54,7 @@ jobs:
         id: download-model-file
         name: Download model file
         with:
-          url: "The model we are using is [tinyllama-1.1b](https://huggingface.co/TheBloke/TinyLlama-1.1B-Chat-v1.0-GGUF/resolve/main/tinyllama-1.1b-chat-v1.0.Q4_K_M.gguf)!"
+          url: ${{ env.LLM_MODEL_URL }}
           target: nitro-node/test/test_assets/
           auto-match: true
           retry-times: 3
@@ -237,119 +72,3 @@ jobs:
         run: |
           cd nitro-node
           make clean test-ci
-
-  #windows-amd64-build:
-  #  runs-on: windows-latest
-  #  steps:
-  #    - name: Clone
-
-  #      id: checkout
-  #      uses: actions/checkout@v4
-  #      with:
-  #        submodules: recursive
-
-  #    - uses: actions/setup-node@v4
-  #      with:
-  #        node-version: 18
-
-  #    - name: Setup VSWhere.exe
-  #      uses: warrenbuckley/Setup-VSWhere@v1
-  #      with:
-  #        version: latest
-  #        silent: true
-  #      env:
-  #        ACTIONS_ALLOW_UNSECURE_COMMANDS: true
-
-  #    - name: Restore cached model file
-  #      id: cache-model-restore
-  #      uses: actions/cache/restore@v4
-  #      with:
-  #        path: |
-  #          nitro-node/test/test_assets/*.gguf
-  #        key: ${{ runner.os }}-model-gguf
-
-  #    - uses: suisei-cn/actions-download-file@v1.4.0
-  #      id: download-model-file
-  #      name: Download model file
-  #      with:
-  #        url: "The model we are using is [tinyllama-1.1b](https://huggingface.co/TheBloke/TinyLlama-1.1B-Chat-v1.0-GGUF/resolve/main/tinyllama-1.1b-chat-v1.0.Q4_K_M.gguf)!"
-  #        target: nitro-node/test/test_assets/
-  #        auto-match: true
-  #        retry-times: 3
-
-  #    - name: Save downloaded model file to cache
-  #      id: cache-model-save
-  #      uses: actions/cache/save@v4
-  #      with:
-  #        path: |
-  #          nitro-node/test/test_assets/*.gguf
-  #        key: ${{ steps.cache-model-restore.outputs.cache-primary-key }}
-
-  #    - name: Run tests
-  #      id: test_nitro_node
-  #      run: |
-  #        cd nitro-node
-  #        make clean test-ci
-
-  #windows-amd64-cuda-build:
-  #  runs-on: windows-cuda-${{ matrix.cuda }}
-  #  strategy:
-  #    matrix:
-  #      cuda: ["12-0", "11-7"]
-
-  #  steps:
-  #    - name: Clone
-  #      id: checkout
-  #      uses: actions/checkout@v4
-  #      with:
-  #        submodules: recursive
-
-  #    - uses: actions/setup-node@v4
-  #      with:
-  #        node-version: 18
-
-  #    - name: actions-setup-cmake
-  #      uses: jwlawson/actions-setup-cmake@v1.14.1
-
-  #    - name: Setup VSWhere.exe
-  #      uses: warrenbuckley/Setup-VSWhere@v1
-  #      with:
-  #        version: latest
-  #        silent: true
-  #      env:
-  #        ACTIONS_ALLOW_UNSECURE_COMMANDS: true
-
-  #    - uses: actions/setup-dotnet@v3
-  #      with:
-  #        dotnet-version: "6.0.x"
-
-  #    - name: Restore cached model file
-  #      id: cache-model-restore
-  #      uses: actions/cache/restore@v4
-  #      with:
-  #        path: |
-  #          nitro-node/test/test_assets/*.gguf
-  #        key: ${{ runner.os }}-model-gguf
-
-  #    - uses: suisei-cn/actions-download-file@v1.4.0
-  #      id: download-model-file
-  #      name: Download model file
-  #      with:
-  #        url: "The model we are using is [tinyllama-1.1b](https://huggingface.co/TheBloke/TinyLlama-1.1B-Chat-v1.0-GGUF/resolve/main/tinyllama-1.1b-chat-v1.0.Q4_K_M.gguf)!"
-  #        target: nitro-node/test/test_assets/
-  #        auto-match: true
-  #        retry-times: 3
-
-  #    - name: Save downloaded model file to cache
-  #      id: cache-model-save
-  #      uses: actions/cache/save@v4
-  #      with:
-  #        path: |
-  #          nitro-node/test/test_assets/*.gguf
-  #        key: ${{ steps.cache-model-restore.outputs.cache-primary-key }}
-
-  #    - name: Run tests
-  #      id: test_nitro_node
-  #      run: |
-  #        cd nitro-node
-  #        make clean test-ci
diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
index c8b799187..4b3b4e5af 100644
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@@ -24,6 +24,7 @@ on:
         "!docs/**",
         "!.gitignore",
         "!README.md",
+        "!nitro-node",
       ]
   pull_request:
     types: [opened, synchronize, reopened]
@@ -44,6 +45,7 @@ on:
         "!docs/**",
         "!.gitignore",
         "!README.md",
+        "!nitro-node",
       ]
   workflow_dispatch:
 
diff --git a/.github/workflows/test-install-nitro-node.yml b/.github/workflows/test-install-nitro-node.yml
index 755551b82..4c526c748 100644
--- a/.github/workflows/test-install-nitro-node.yml
+++ b/.github/workflows/test-install-nitro-node.yml
@@ -50,45 +50,19 @@ jobs:
           path: nitro-node/janhq-nitro-node.tgz
           if-no-files-found: error
 
-  ubuntu-install:
-    runs-on: ubuntu-latest
-    needs: [linux-pack-tarball]
-    if: always() && needs.linux-pack-tarball.result == 'success'
-    steps:
-      - name: Clone
-        id: checkout
-        uses: actions/checkout@v4
-        with:
-          submodules: recursive
-
-      - uses: actions/setup-node@v4
-        with:
-          node-version: 18
-
-      - name: Download prebuilt tarball
-        uses: actions/download-artifact@master
-        with:
-          name: janhq-nitro-node
-          path: .github/scripts/
-
-      - name: List tarball content
-        id: tar-tf
-        run: |
-          cd .github
-          cd scripts
-          tar tf janhq-nitro-node.tgz
-
-      - name: Run tests
-        id: test_install_nitro_node
-        env:
-          NITRO_NODE_PKG: ${{ github.workspace }}/.github/scripts/janhq-nitro-node.tgz
-        run: |
-          cd .github
-          cd scripts
-          node e2e-test-install-nitro-node.js
-
-  macOS-install:
-    runs-on: macos-latest
+  install:
+    runs-on: ${{ matrix.os }}
+    strategy:
+      matrix:
+        os:
+          - ubuntu-latest
+          - ubuntu-18-04-cuda-12-0
+          - ubuntu-18-04-cuda-11-7
+          - macos-silicon
+          - macos-latest
+          - windows-latest
+          - windows-cuda-12-0
+          - windows-cuda-11-7
     needs: [linux-pack-tarball]
     if: always() && needs.linux-pack-tarball.result == 'success'
     steps:
@@ -123,41 +97,3 @@ jobs:
           cd .github
           cd scripts
           node e2e-test-install-nitro-node.js
-
-  windows-install:
-    runs-on: windows-latest
-    needs: [linux-pack-tarball]
-    if: always() && needs.linux-pack-tarball.result == 'success'
-    steps:
-      - name: Clone
-
-        id: checkout
-        uses: actions/checkout@v4
-        with:
-          submodules: recursive
-
-      - uses: actions/setup-node@v4
-        with:
-          node-version: 18
-
-      - name: Download prebuilt tarball
-        uses: actions/download-artifact@master
-        with:
-          name: janhq-nitro-node
-          path: .github/scripts/
-
-      - name: List tarball content
-        id: tar-tf
-        run: |
-          cd .github
-          cd scripts
-          tar tf janhq-nitro-node.tgz
-
-      - name: Run tests
-        id: test_install_nitro_node
-        env:
-          NITRO_NODE_PKG: ${{ github.workspace }}\.github\scripts\janhq-nitro-node.tgz
-        run: |
-          cd .github
-          cd scripts
-          node e2e-test-install-nitro-node.js