CUDA GPU Support - External Provider Binaries

.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.1.12
+current_version = 0.1.13
 commit = True
 tag = True
 tag_name = v{new_version}

.dockerignore

@@ -0,0 +1,55 @@
+# Dependencies
+node_modules/
+__pycache__/
+*.pyc
+*.pyo
+*.pyd
+.Python
+env/
+venv/
+
+# Build outputs
+build/
+*.egg-info/
+.eggs/
+target/
+
+# Keep web/dist for the Docker image
+!web/dist
+
+# Development
+.git/
+.github/
+.vscode/
+.idea/
+*.swp
+*.swo
+*~
+
+# Data and logs
+data/
+*.log
+*.sqlite
+*.db
+
+# OS
+.DS_Store
+Thumbs.db
+
+# Documentation
+docs/
+landing/
+mlx-test/
+
+# Test files
+*.test.ts
+*.test.tsx
+*.spec.ts
+*.spec.tsx
+
+# Keep these out
+.env
+.env.local
+*.pem
+*.key
+credentials.json

.github/workflows/release.yml

@@ -6,26 +6,174 @@ on:
     tags:
       - "v*"
 
+env:
+  PROVIDER_VERSION: "1.0.0"
+
 jobs:
+  # ============================================
+  # Build TTS Providers (uploaded to R2, not GitHub)
+  # ============================================
+  build-providers:
+    runs-on: ${{ matrix.platform }}
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          # PyTorch CPU provider (Windows)
+          - platform: "windows-latest"
+            provider: "pytorch-cpu"
+            python-version: "3.12"
+          # PyTorch CUDA provider (Windows) - large binary, uploaded to R2
+          - platform: "windows-latest"
+            provider: "pytorch-cuda"
+            python-version: "3.12"
+          # PyTorch CPU provider (Linux)
+          - platform: "ubuntu-22.04"
+            provider: "pytorch-cpu"
+            python-version: "3.12"
+          # PyTorch CUDA provider (Linux) - large binary, uploaded to R2
+          - platform: "ubuntu-22.04"
+            provider: "pytorch-cuda"
+            python-version: "3.12"
+          # PyTorch CPU provider (macOS Apple Silicon)
+          - platform: "macos-latest"
+            provider: "pytorch-cpu"
+            python-version: "3.12"
+          # PyTorch CPU provider (macOS Intel)
+          - platform: "macos-15-intel"
+            provider: "pytorch-cpu"
+            python-version: "3.12"
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Install dependencies (ubuntu only)
+        if: matrix.platform == 'ubuntu-22.04'
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y llvm-dev
+
+      - name: Setup Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+          cache: "pip"
+
+      - name: Install CPU-only torch (Linux)
+        if: matrix.provider == 'pytorch-cpu' && matrix.platform == 'ubuntu-22.04'
+        run: |
+          python -m pip install --upgrade pip
+          pip install pyinstaller
+          pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu
+          pip install -r providers/pytorch-cpu/requirements.txt
+          pip install -r backend/requirements.txt
+
+      - name: Install Python dependencies (CPU - non-Linux)
+        if: matrix.provider == 'pytorch-cpu' && matrix.platform != 'ubuntu-22.04'
+        run: |
+          python -m pip install --upgrade pip
+          pip install pyinstaller
+          pip install -r providers/pytorch-cpu/requirements.txt
+          pip install -r backend/requirements.txt
+
+      - name: Install Python dependencies (CUDA)
+        if: matrix.provider == 'pytorch-cuda'
+        run: |
+          python -m pip install --upgrade pip
+          pip install pyinstaller
+          pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
+          pip install -r providers/pytorch-cuda/requirements.txt
+          pip install -r backend/requirements.txt
+
+      - name: Build provider binary
+        shell: bash
+        run: |
+          cd providers/${{ matrix.provider }}
+          python build.py
+
+      - name: Package provider for distribution
+        shell: bash
+        run: |
+          cd providers/${{ matrix.provider }}/dist
+
+          # Add platform suffix for archive name
+          if [ "${{ matrix.platform }}" == "windows-latest" ]; then
+            ARCHIVE_NAME="tts-provider-${{ matrix.provider }}-windows.zip"
+            # On Windows, zip the directory
+            powershell Compress-Archive -Path "tts-provider-${{ matrix.provider }}/*" -DestinationPath "$ARCHIVE_NAME"
+          elif [ "${{ matrix.platform }}" == "macos-latest" ]; then
+            ARCHIVE_NAME="tts-provider-${{ matrix.provider }}-macos-arm64.tar.gz"
+            tar -czf "$ARCHIVE_NAME" tts-provider-${{ matrix.provider }}/
+          elif [ "${{ matrix.platform }}" == "macos-15-intel" ]; then
+            ARCHIVE_NAME="tts-provider-${{ matrix.provider }}-macos-x64.tar.gz"
+            tar -czf "$ARCHIVE_NAME" tts-provider-${{ matrix.provider }}/
+          else
+            ARCHIVE_NAME="tts-provider-${{ matrix.provider }}-linux.tar.gz"
+            tar -czf "$ARCHIVE_NAME" tts-provider-${{ matrix.provider }}/
+          fi
+
+          echo "Created archive: $ARCHIVE_NAME"
+          ls -lh "$ARCHIVE_NAME"
+
+      - name: Upload provider to R2
+        shell: bash
+        env:
+          R2_ACCESS_KEY_ID: ${{ secrets.R2_ACCESS_KEY_ID }}
+          R2_SECRET_ACCESS_KEY: ${{ secrets.R2_SECRET_ACCESS_KEY }}
+          R2_ENDPOINT: ${{ secrets.R2_ENDPOINT }}
+        run: |
+          # Install AWS CLI (compatible with R2)
+          pip install awscli
+
+          # Configure AWS CLI for R2
+          aws configure set aws_access_key_id $R2_ACCESS_KEY_ID
+          aws configure set aws_secret_access_key $R2_SECRET_ACCESS_KEY
+          aws configure set region auto
+
+          # Determine archive name based on platform
+          if [ "${{ matrix.platform }}" == "windows-latest" ]; then
+            ARCHIVE_NAME="tts-provider-${{ matrix.provider }}-windows.zip"
+          elif [ "${{ matrix.platform }}" == "macos-latest" ]; then
+            ARCHIVE_NAME="tts-provider-${{ matrix.provider }}-macos-arm64.tar.gz"
+          elif [ "${{ matrix.platform }}" == "macos-15-intel" ]; then
+            ARCHIVE_NAME="tts-provider-${{ matrix.provider }}-macos-x64.tar.gz"
+          else
+            ARCHIVE_NAME="tts-provider-${{ matrix.provider }}-linux.tar.gz"
+          fi
+
+          # Upload to R2 (bucket: voicebox)
+          aws s3 cp "providers/${{ matrix.provider }}/dist/$ARCHIVE_NAME" \
+            "s3://voicebox/providers/v${{ env.PROVIDER_VERSION }}/$ARCHIVE_NAME" \
+            --endpoint-url "$R2_ENDPOINT"
+
+          echo "Uploaded $ARCHIVE_NAME to R2"
+
+  # ============================================
+  # Build Main App (without bundled TTS on Win/Linux)
+  # ============================================
   release:
     permissions:
       contents: write
     strategy:
       fail-fast: false
       matrix:
         include:
+          # macOS Apple Silicon - MLX bundled (works out of the box)
           - platform: "macos-latest"
             args: "--target aarch64-apple-darwin"
             python-version: "3.12"
             backend: "mlx"
+          # macOS Intel - PyTorch bundled (smaller user base, keep simple)
           - platform: "macos-15-intel"
             args: "--target x86_64-apple-darwin"
             python-version: "3.12"
             backend: "pytorch"
-          # - platform: 'ubuntu-22.04'
-          #   args: ''
-          #   python-version: '3.12'
-          #   backend: 'pytorch'
+          # Linux - No TTS bundled, providers downloaded separately
+          - platform: "ubuntu-22.04"
+            args: ""
+            python-version: "3.12"
+            backend: "none"
+          # Windows - PyTorch CPU bundled (works out of the box)
           - platform: "windows-latest"
             args: ""
             python-version: "3.12"
@@ -40,7 +188,7 @@ jobs:
         if: matrix.platform == 'ubuntu-22.04'
         run: |
           sudo apt-get update
-          sudo apt-get install -y libwebkit2gtk-4.1-dev libappindicator3-dev librsvg2-dev patchelf llvm-dev
+          sudo apt-get install -y libwebkit2gtk-4.1-dev libappindicator3-dev librsvg2-dev patchelf llvm-dev libasound2-dev
 
       - name: Install LLVM (macOS)
         if: matrix.platform == 'macos-latest' || matrix.platform == 'macos-15-intel'
@@ -55,23 +203,27 @@ jobs:
           python-version: ${{ matrix.python-version }}
           cache: "pip"
 
-      - name: Install Python dependencies
+      - name: Install Python dependencies (with TTS)
+        if: matrix.backend != 'none'
         run: |
           python -m pip install --upgrade pip
           pip install pyinstaller
           pip install -r backend/requirements.txt
 
+      - name: Install Python dependencies (without TTS)
+        if: matrix.backend == 'none'
+        run: |
+          python -m pip install --upgrade pip
+          pip install pyinstaller
+          # Install base requirements without PyTorch/Qwen-TTS
+          pip install fastapi uvicorn sqlalchemy librosa soundfile numpy httpx
+          pip install huggingface_hub  # For Whisper downloads
+
       - name: Install MLX dependencies (Apple Silicon only)
         if: matrix.backend == 'mlx'
         run: |
           pip install -r backend/requirements-mlx.txt
 
-      # - name: Install PyTorch with CUDA (Windows only)
-      #   if: matrix.platform == 'windows-latest'
-      #   run: |
-      #     pip install torch --index-url https://download.pytorch.org/whl/cu121 --force-reinstall --no-deps
-      #     pip install torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
-
       - name: Build Python server (Linux/macOS)
         if: matrix.platform != 'windows-latest'
         run: |
@@ -148,13 +300,84 @@ jobs:
             See the assets below to download and install this version.
 
             ### Installation
-            - **macOS (Apple Silicon)**: Download the `aarch64.dmg` file - uses MLX for fast native inference
+            - **macOS (Apple Silicon)**: Download the `aarch64.dmg` file - uses MLX for fast native inference (works out of the box)
             - **macOS (Intel)**: Download the `x64.dmg` file - uses PyTorch
-            - **Windows**: Download the `.msi` installer
-            - **Linux**: Download the `.AppImage` or `.deb` package
+            - **Windows**: Download the `.msi` installer - requires downloading a TTS provider on first use
+            - **Linux**: Download the `.AppImage` or `.deb` package - requires downloading a TTS provider on first use
+
+            ### TTS Providers
+            Windows and Linux users will be prompted to download a TTS provider on first launch:
+            - **Windows**: PyTorch CPU (~300MB) or PyTorch CUDA (~2.4GB for NVIDIA GPUs)
+            - **Linux**: PyTorch CUDA (~2.4GB) - requires NVIDIA GPU
 
             The app includes automatic updates - future updates will be installed automatically.
           releaseDraft: true
           prerelease: false
           args: ${{ matrix.args }}
           includeUpdaterJson: true
+
+  # ============================================
+  # Build and Push Docker Images
+  # ============================================
+  docker:
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      packages: write
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Login to GitHub Container Registry
+        uses: docker/login-action@v3
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Setup Bun
+        uses: oven-sh/setup-bun@v2
+
+      - name: Install dependencies and build web UI
+        run: |
+          bun install
+          cd web
+          bun run build
+
+      - name: Extract version from tag
+        id: version
+        run: |
+          if [[ $GITHUB_REF == refs/tags/v* ]]; then
+            VERSION=${GITHUB_REF#refs/tags/v}
+          else
+            VERSION="dev"
+          fi
+          echo "version=$VERSION" >> $GITHUB_OUTPUT
+
+      - name: Build and push CPU image
+        uses: docker/build-push-action@v5
+        with:
+          context: .
+          file: ./Dockerfile
+          platforms: linux/amd64,linux/arm64
+          push: true
+          tags: |
+            ghcr.io/jamiepine/voicebox:latest
+            ghcr.io/jamiepine/voicebox:${{ steps.version.outputs.version }}
+          cache-from: type=gha
+          cache-to: type=gha,mode=max
+
+      - name: Build and push CUDA image
+        uses: docker/build-push-action@v5
+        with:
+          context: .
+          file: ./Dockerfile.cuda
+          platforms: linux/amd64
+          push: true
+          tags: |
+            ghcr.io/jamiepine/voicebox:latest-cuda
+            ghcr.io/jamiepine/voicebox:${{ steps.version.outputs.version }}-cuda
+          cache-from: type=gha
+          cache-to: type=gha,mode=max

.gitignore

@@ -15,6 +15,7 @@ dist/
 build/
 *.egg-info/
 *.egg
+*.spec
 target/
 *.app
 *.dmg