chore: bump sidecar version to 1.0.15 [skip ci]

Engines now read user.name from the config object at transcription time
instead of caching it at init, so name changes take effect immediately.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

.gitea/workflows/build-app-linux.yml

@@ -13,10 +13,11 @@ jobs:
     runs-on: ubuntu-latest
     env:
       NODE_VERSION: "20"
-      RELEASE_TAG: ${{ inputs.tag }}
+      RELEASE_TAG: "${{ inputs.tag }}"
     steps:
       - name: Show tag
-        run: echo "Building for tag: ${RELEASE_TAG}"
+        run: |
+          echo "Building for tag: ${RELEASE_TAG}"
 
       - uses: actions/checkout@v4
         with:

.gitea/workflows/build-app-macos.yml

@@ -13,10 +13,11 @@ jobs:
     runs-on: macos-latest
     env:
       NODE_VERSION: "20"
-      RELEASE_TAG: ${{ inputs.tag }}
+      RELEASE_TAG: "${{ inputs.tag }}"
     steps:
       - name: Show tag
-        run: echo "Building for tag: ${RELEASE_TAG}"
+        run: |
+          echo "Building for tag: ${RELEASE_TAG}"
 
       - uses: actions/checkout@v4
         with:

.gitea/workflows/build-app-windows.yml

@@ -15,7 +15,7 @@ jobs:
     name: Build App (Windows)
     runs-on: windows-latest
     env:
-      RELEASE_TAG: ${{ inputs.tag }}
+      RELEASE_TAG: "${{ inputs.tag }}"
     steps:
       - name: Show tag
         shell: powershell
@@ -46,8 +46,45 @@ jobs:
         shell: powershell
         run: npm ci
 
+      - name: Setup Azure Artifact Signing
+        shell: powershell
+        env:
+          AZURE_CLIENT_ID: ${{ secrets.AZURE_CLIENT_ID }}
+          AZURE_SIGNING_ENDPOINT: ${{ secrets.AZURE_SIGNING_ENDPOINT }}
+          AZURE_SIGNING_ACCOUNT: ${{ secrets.AZURE_SIGNING_ACCOUNT }}
+          AZURE_CERT_PROFILE: ${{ secrets.AZURE_CERT_PROFILE }}
+        run: |
+          if (-not $env:AZURE_CLIENT_ID) {
+            Write-Host "No Azure signing secrets configured, skipping code signing setup"
+            return
+          }
+
+          Write-Host "Setting up Azure Artifact Signing..."
+
+          # Install Artifact Signing client tools
+          nuget install Microsoft.ArtifactSigning.Client -x -OutputDirectory .\signing-tools
+          $dlibPath = (Resolve-Path ".\signing-tools\Microsoft.ArtifactSigning.Client*\bin\x64\Azure.CodeSigning.Dlib.dll").Path
+
+          # Write metadata.json
+          @{
+            Endpoint = $env:AZURE_SIGNING_ENDPOINT
+            CodeSigningAccountName = $env:AZURE_SIGNING_ACCOUNT
+            CertificateProfileName = $env:AZURE_CERT_PROFILE
+          } | ConvertTo-Json | Out-File -Encoding UTF8 metadata.json
+          $metadataPath = (Resolve-Path "metadata.json").Path
+
+          # Inject signCommand into tauri.conf.json for this build
+          $conf = Get-Content src-tauri\tauri.conf.json -Raw | ConvertFrom-Json
+          $signCmd = "signtool.exe sign /v /fd SHA256 /tr http://timestamp.acs.microsoft.com /td SHA256 /dlib `"$dlibPath`" /dmdf `"$metadataPath`" %1"
+          $conf.bundle.windows | Add-Member -NotePropertyName "signCommand" -NotePropertyValue $signCmd -Force
+          $conf | ConvertTo-Json -Depth 10 | Set-Content src-tauri\tauri.conf.json -Encoding UTF8
+
       - name: Build Tauri app
         shell: powershell
+        env:
+          AZURE_CLIENT_ID: ${{ secrets.AZURE_CLIENT_ID }}
+          AZURE_CLIENT_SECRET: ${{ secrets.AZURE_CLIENT_SECRET }}
+          AZURE_TENANT_ID: ${{ secrets.AZURE_TENANT_ID }}
         run: npm run tauri build
 
       - name: Upload to release

.gitea/workflows/build-sidecar-cloud.yml

@@ -0,0 +1,229 @@
+name: Build Sidecar (Cloud)
+
+on:
+  workflow_dispatch:
+    inputs:
+      tag:
+        description: 'Sidecar release tag to build (e.g. sidecar-v1.0.5)'
+        required: true
+
+jobs:
+  build-cloud-linux:
+    name: Build Cloud Sidecar (Linux)
+    runs-on: ubuntu-latest
+    env:
+      PYTHON_VERSION: "3.11"
+      RELEASE_TAG: "${{ inputs.tag }}"
+    steps:
+      - name: Show tag
+        run: |
+          echo "Building cloud sidecar for tag ${RELEASE_TAG}"
+
+      - uses: actions/checkout@v4
+        with:
+          ref: ${{ inputs.tag }}
+
+      - name: Install uv
+        run: |
+          curl -LsSf https://astral.sh/uv/install.sh | sh
+          echo "$HOME/.local/bin" >> $GITHUB_PATH
+
+      - name: Set up Python
+        run: uv python install ${{ env.PYTHON_VERSION }}
+
+      - name: Install system dependencies
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y portaudio19-dev
+
+      - name: Build cloud sidecar
+        env:
+          UV_NO_SOURCES: "1"
+        run: |
+          uv venv
+          uv pip install pyinstaller numpy sounddevice fastapi uvicorn websockets pydantic requests pyyaml packaging
+          .venv/bin/pyinstaller local-transcription-cloud.spec
+
+      - name: Package
+        run: |
+          cd dist/local-transcription-backend && zip -r ../../sidecar-linux-x86_64-cloud.zip .
+
+      - name: Upload to release
+        env:
+          BUILD_TOKEN: ${{ secrets.BUILD_TOKEN }}
+        run: |
+          sudo apt-get install -y jq
+          REPO_API="${GITHUB_SERVER_URL}/api/v1/repos/${GITHUB_REPOSITORY}"
+          TAG="${RELEASE_TAG}"
+
+          for i in $(seq 1 30); do
+            RELEASE_ID=$(curl -s -H "Authorization: token ${BUILD_TOKEN}" \
+              "${REPO_API}/releases/tags/${TAG}" | jq -r '.id // empty')
+            if [ -n "${RELEASE_ID}" ] && [ "${RELEASE_ID}" != "null" ]; then
+              echo "Found release ${TAG} (ID: ${RELEASE_ID})"
+              break
+            fi
+            echo "Attempt ${i}/30: waiting for release..."
+            sleep 10
+          done
+
+          if [ -z "${RELEASE_ID}" ] || [ "${RELEASE_ID}" = "null" ]; then
+            echo "ERROR: Release not found"; exit 1
+          fi
+
+          for file in sidecar-*-cloud.zip; do
+            filename=$(basename "$file")
+            ASSET_ID=$(curl -s -H "Authorization: token ${BUILD_TOKEN}" \
+              "${REPO_API}/releases/${RELEASE_ID}/assets" | jq -r ".[] | select(.name == \"${filename}\") | .id // empty")
+            [ -n "${ASSET_ID}" ] && curl -s -X DELETE -H "Authorization: token ${BUILD_TOKEN}" "${REPO_API}/releases/${RELEASE_ID}/assets/${ASSET_ID}"
+            curl -s -o /dev/null -w "Upload ${filename}: HTTP %{http_code}\n" -X POST \
+              -H "Authorization: token ${BUILD_TOKEN}" -H "Content-Type: application/octet-stream" \
+              -T "$file" "${REPO_API}/releases/${RELEASE_ID}/assets?name=${filename}"
+          done
+
+  build-cloud-windows:
+    name: Build Cloud Sidecar (Windows)
+    runs-on: windows-latest
+    env:
+      PYTHON_VERSION: "3.11"
+      RELEASE_TAG: "${{ inputs.tag }}"
+    steps:
+      - name: Show tag
+        shell: powershell
+        run: Write-Host "Building cloud sidecar for tag $env:RELEASE_TAG"
+
+      - uses: actions/checkout@v4
+        with:
+          ref: ${{ inputs.tag }}
+
+      - name: Install uv
+        shell: powershell
+        run: |
+          if (Get-Command uv -ErrorAction SilentlyContinue) {
+            Write-Host "uv already installed"
+          } else {
+            irm https://astral.sh/uv/install.ps1 | iex
+            $uvPaths = @("$env:USERPROFILE\.local\bin", "$env:USERPROFILE\.cargo\bin", "$env:LOCALAPPDATA\uv\bin")
+            foreach ($p in $uvPaths) { if (Test-Path $p) { echo $p | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append } }
+          }
+
+      - name: Set up Python
+        shell: powershell
+        run: uv python install ${{ env.PYTHON_VERSION }}
+
+      - name: Build cloud sidecar
+        shell: powershell
+        env:
+          UV_NO_SOURCES: "1"
+        run: |
+          uv venv
+          uv pip install pyinstaller numpy sounddevice fastapi uvicorn websockets pydantic requests pyyaml packaging
+          .venv\Scripts\pyinstaller.exe local-transcription-cloud.spec
+
+      - name: Package
+        shell: powershell
+        run: |
+          if (-not (Get-Command 7z -ErrorAction SilentlyContinue)) { choco install 7zip -y }
+          7z a -tzip -mx=5 sidecar-windows-x86_64-cloud.zip .\dist\local-transcription-backend\*
+
+      - name: Upload to release
+        shell: powershell
+        env:
+          BUILD_TOKEN: ${{ secrets.BUILD_TOKEN }}
+        run: |
+          $REPO_API = "${{ github.server_url }}/api/v1/repos/${{ github.repository }}"
+          $Headers = @{ "Authorization" = "token $env:BUILD_TOKEN" }
+          $TAG = $env:RELEASE_TAG
+
+          $RELEASE_ID = $null
+          for ($i = 1; $i -le 30; $i++) {
+            try {
+              $release = Invoke-RestMethod -Uri "$REPO_API/releases/tags/$TAG" -Headers $Headers -ErrorAction Stop
+              $RELEASE_ID = $release.id
+              if ($RELEASE_ID) { Write-Host "Found release $TAG (ID: $RELEASE_ID)"; break }
+            } catch {}
+            Write-Host "Attempt ${i}/30: waiting..."; Start-Sleep -Seconds 10
+          }
+          if (-not $RELEASE_ID) { Write-Host "ERROR: Release not found"; exit 1 }
+
+          Get-ChildItem -Path . -Filter "sidecar-*-cloud.zip" | ForEach-Object {
+            $fn = $_.Name; $enc = [System.Uri]::EscapeDataString($fn)
+            try {
+              $assets = Invoke-RestMethod -Uri "$REPO_API/releases/$RELEASE_ID/assets" -Headers $Headers
+              $existing = $assets | Where-Object { $_.name -eq $fn }
+              if ($existing) { Invoke-RestMethod -Uri "$REPO_API/releases/$RELEASE_ID/assets/$($existing.id)" -Method Delete -Headers $Headers }
+            } catch {}
+            curl.exe --fail -s -X POST -H "Authorization: token $env:BUILD_TOKEN" -H "Content-Type: application/octet-stream" -T "$($_.FullName)" "$REPO_API/releases/$RELEASE_ID/assets?name=$enc"
+            Write-Host "Uploaded $fn"
+          }
+
+  build-cloud-macos:
+    name: Build Cloud Sidecar (macOS)
+    runs-on: macos-latest
+    env:
+      PYTHON_VERSION: "3.11"
+      RELEASE_TAG: "${{ inputs.tag }}"
+    steps:
+      - name: Show tag
+        run: |
+          echo "Building cloud sidecar for tag ${RELEASE_TAG}"
+
+      - uses: actions/checkout@v4
+        with:
+          ref: ${{ inputs.tag }}
+
+      - name: Install uv
+        run: |
+          curl -LsSf https://astral.sh/uv/install.sh | sh
+          echo "$HOME/.local/bin" >> $GITHUB_PATH
+
+      - name: Set up Python
+        run: uv python install ${{ env.PYTHON_VERSION }}
+
+      - name: Install system dependencies
+        run: brew install portaudio
+
+      - name: Build cloud sidecar
+        env:
+          UV_NO_SOURCES: "1"
+        run: |
+          uv venv
+          uv pip install pyinstaller numpy sounddevice fastapi uvicorn websockets pydantic requests pyyaml packaging
+          .venv/bin/pyinstaller local-transcription-cloud.spec
+
+      - name: Package
+        run: |
+          cd dist/local-transcription-backend && zip -r ../../sidecar-macos-aarch64-cloud.zip .
+
+      - name: Upload to release
+        env:
+          BUILD_TOKEN: ${{ secrets.BUILD_TOKEN }}
+        run: |
+          which jq || brew install jq
+          REPO_API="${GITHUB_SERVER_URL}/api/v1/repos/${GITHUB_REPOSITORY}"
+          TAG="${RELEASE_TAG}"
+
+          for i in $(seq 1 30); do
+            RELEASE_ID=$(curl -s -H "Authorization: token ${BUILD_TOKEN}" \
+              "${REPO_API}/releases/tags/${TAG}" | jq -r '.id // empty')
+            if [ -n "${RELEASE_ID}" ] && [ "${RELEASE_ID}" != "null" ]; then
+              echo "Found release ${TAG} (ID: ${RELEASE_ID})"
+              break
+            fi
+            echo "Attempt ${i}/30: waiting for release..."
+            sleep 10
+          done
+
+          if [ -z "${RELEASE_ID}" ] || [ "${RELEASE_ID}" = "null" ]; then
+            echo "ERROR: Release not found"; exit 1
+          fi
+
+          for file in sidecar-*-cloud.zip; do
+            filename=$(basename "$file")
+            ASSET_ID=$(curl -s -H "Authorization: token ${BUILD_TOKEN}" \
+              "${REPO_API}/releases/${RELEASE_ID}/assets" | jq -r ".[] | select(.name == \"${filename}\") | .id // empty")
+            [ -n "${ASSET_ID}" ] && curl -s -X DELETE -H "Authorization: token ${BUILD_TOKEN}" "${REPO_API}/releases/${RELEASE_ID}/assets/${ASSET_ID}"
+            curl -s -o /dev/null -w "Upload ${filename}: HTTP %{http_code}\n" -X POST \
+              -H "Authorization: token ${BUILD_TOKEN}" -H "Content-Type: application/octet-stream" \
+              -T "$file" "${REPO_API}/releases/${RELEASE_ID}/assets?name=${filename}"
+          done

.gitea/workflows/build-sidecar-linux.yml

@@ -13,10 +13,11 @@ jobs:
     runs-on: ubuntu-latest
     env:
       PYTHON_VERSION: "3.11"
-      RELEASE_TAG: ${{ inputs.tag }}
+      RELEASE_TAG: "${{ inputs.tag }}"
     steps:
       - name: Show tag
-        run: echo "Building for tag: ${RELEASE_TAG}"
+        run: |
+          echo "Building for tag: ${RELEASE_TAG}"
 
       - uses: actions/checkout@v4
         with:
@@ -39,26 +40,17 @@ jobs:
           sudo apt-get update
           sudo apt-get install -y portaudio19-dev
 
-      - name: Build sidecar (CUDA)
-        run: |
-          uv sync --frozen || uv sync
-          uv run pyinstaller local-transcription-headless.spec
-
-      - name: Package sidecar (CUDA)
-        run: |
-          cd dist/local-transcription-backend && zip -r ../../sidecar-linux-x86_64-cuda.zip .
-
       - name: Build sidecar (CPU)
         run: |
-          rm -rf dist/local-transcription-backend build/
+          uv sync --no-sources
+          # PyPI's default torch on Linux includes CUDA (~800MB).
+          # Replace with CPU-only torch from the dedicated index.
           uv pip install torch torchaudio --index-url https://download.pytorch.org/whl/cpu --force-reinstall
-          # Run pyinstaller directly from venv to prevent uv run from
-          # re-resolving torch back to the CUDA version via pyproject.toml sources
           .venv/bin/pyinstaller local-transcription-headless.spec
 
       - name: Package sidecar (CPU)
         run: |
-          cd dist/local-transcription-backend && zip -r ../../sidecar-linux-x86_64-cpu.zip .
+          cd dist/local-transcription-backend && zip -9 -r ../../sidecar-linux-x86_64-cpu.zip .
 
       - name: Upload to sidecar release
         env:

.gitea/workflows/build-sidecar-macos.yml

@@ -13,10 +13,11 @@ jobs:
     runs-on: macos-latest
     env:
       PYTHON_VERSION: "3.11"
-      RELEASE_TAG: ${{ inputs.tag }}
+      RELEASE_TAG: "${{ inputs.tag }}"
     steps:
       - name: Show tag
-        run: echo "Building for tag: ${RELEASE_TAG}"
+        run: |
+          echo "Building for tag: ${RELEASE_TAG}"
 
       - uses: actions/checkout@v4
         with:

.gitea/workflows/build-sidecar-windows.yml

@@ -13,7 +13,7 @@ jobs:
     runs-on: windows-latest
     env:
       PYTHON_VERSION: "3.11"
-      RELEASE_TAG: ${{ inputs.tag }}
+      RELEASE_TAG: "${{ inputs.tag }}"
     steps:
       - name: Show tag
         shell: powershell
@@ -54,29 +54,19 @@ jobs:
             choco install 7zip -y
           }
 
-      - name: Build sidecar (CUDA)
-        shell: powershell
-        run: |
-          uv sync --frozen
-          if ($LASTEXITCODE -ne 0) { uv sync }
-          uv run pyinstaller local-transcription-headless.spec
-
-      - name: Package sidecar (CUDA)
-        shell: powershell
-        run: |
-          7z a -tzip -mx=5 sidecar-windows-x86_64-cuda.zip .\dist\local-transcription-backend\*
-
       - name: Build sidecar (CPU)
         shell: powershell
         run: |
-          Remove-Item -Recurse -Force dist\local-transcription-backend, build -ErrorAction SilentlyContinue
+          $env:UV_NO_SOURCES = "1"
+          uv sync
+          # PyPI's default torch includes CUDA. Replace with CPU-only.
           uv pip install torch torchaudio --index-url https://download.pytorch.org/whl/cpu --force-reinstall
           .venv\Scripts\pyinstaller.exe local-transcription-headless.spec
 
       - name: Package sidecar (CPU)
         shell: powershell
         run: |
-          7z a -tzip -mx=5 sidecar-windows-x86_64-cpu.zip .\dist\local-transcription-backend\*
+          7z a -tzip -mx=9 sidecar-windows-x86_64-cpu.zip .\dist\local-transcription-backend\*
 
       - name: Upload to sidecar release
         shell: powershell

.gitea/workflows/cleanup-releases.yml

@@ -0,0 +1,102 @@
+name: Cleanup Old Releases
+
+on:
+  workflow_dispatch:
+    inputs:
+      keep_app_releases:
+        description: 'Number of app releases to keep'
+        required: false
+        default: '3'
+      keep_sidecar_releases:
+        description: 'Number of sidecar releases to keep'
+        required: false
+        default: '2'
+      dry_run:
+        description: 'Dry run (show what would be deleted without deleting)'
+        required: false
+        default: 'true'
+
+jobs:
+  cleanup:
+    name: Cleanup Old Releases
+    runs-on: ubuntu-latest
+    steps:
+      - name: Cleanup releases
+        env:
+          BUILD_TOKEN: ${{ secrets.BUILD_TOKEN }}
+        run: |
+          REPO_API="${GITHUB_SERVER_URL}/api/v1/repos/${GITHUB_REPOSITORY}"
+          KEEP_APP="${{ inputs.keep_app_releases }}"
+          KEEP_SIDECAR="${{ inputs.keep_sidecar_releases }}"
+          DRY_RUN="${{ inputs.dry_run }}"
+
+          echo "=== Cleanup Configuration ==="
+          echo "Keep app releases: ${KEEP_APP}"
+          echo "Keep sidecar releases: ${KEEP_SIDECAR}"
+          echo "Dry run: ${DRY_RUN}"
+          echo ""
+
+          # Fetch all releases
+          ALL_RELEASES=$(curl -s -H "Authorization: token ${BUILD_TOKEN}" \
+            "${REPO_API}/releases?limit=50")
+
+          # ── App releases (v* tags, not sidecar-v*) ──
+          echo "=== App Releases ==="
+          APP_RELEASES=$(echo "$ALL_RELEASES" | jq -c '[.[] | select(.tag_name | startswith("v")) | select(.tag_name | startswith("sidecar") | not)]')
+          APP_TOTAL=$(echo "$APP_RELEASES" | jq 'length')
+          echo "Found ${APP_TOTAL} app releases, keeping ${KEEP_APP}"
+
+          if [ "$APP_TOTAL" -gt "$KEEP_APP" ]; then
+            echo "$APP_RELEASES" | jq -c ".[$KEEP_APP:][]" | while read -r release; do
+              ID=$(echo "$release" | jq -r '.id')
+              TAG=$(echo "$release" | jq -r '.tag_name')
+              SIZE=$(echo "$release" | jq '[.assets[]?.size // 0] | add // 0')
+              SIZE_MB=$(echo "scale=1; $SIZE / 1048576" | bc 2>/dev/null || echo "?")
+
+              # Protect v1.4.0 (last pre-Tauri release)
+              if [ "$TAG" = "v1.4.0" ]; then
+                echo "  PROTECT ${TAG} (${SIZE_MB} MB)"
+                continue
+              fi
+
+              if [ "$DRY_RUN" = "true" ]; then
+                echo "  WOULD DELETE ${TAG} (ID: ${ID}, ${SIZE_MB} MB)"
+              else
+                echo "  DELETING ${TAG} (ID: ${ID}, ${SIZE_MB} MB)..."
+                curl -s -X DELETE -H "Authorization: token ${BUILD_TOKEN}" \
+                  "${REPO_API}/releases/${ID}"
+              fi
+            done
+          else
+            echo "  Nothing to clean up"
+          fi
+
+          echo ""
+
+          # ── Sidecar releases (sidecar-v* tags) ──
+          echo "=== Sidecar Releases ==="
+          SIDECAR_RELEASES=$(echo "$ALL_RELEASES" | jq -c '[.[] | select(.tag_name | startswith("sidecar-v"))]')
+          SIDECAR_TOTAL=$(echo "$SIDECAR_RELEASES" | jq 'length')
+          echo "Found ${SIDECAR_TOTAL} sidecar releases, keeping ${KEEP_SIDECAR}"
+
+          if [ "$SIDECAR_TOTAL" -gt "$KEEP_SIDECAR" ]; then
+            echo "$SIDECAR_RELEASES" | jq -c ".[$KEEP_SIDECAR:][]" | while read -r release; do
+              ID=$(echo "$release" | jq -r '.id')
+              TAG=$(echo "$release" | jq -r '.tag_name')
+              SIZE=$(echo "$release" | jq '[.assets[]?.size // 0] | add // 0')
+              SIZE_MB=$(echo "scale=1; $SIZE / 1048576" | bc 2>/dev/null || echo "?")
+
+              if [ "$DRY_RUN" = "true" ]; then
+                echo "  WOULD DELETE ${TAG} (ID: ${ID}, ${SIZE_MB} MB)"
+              else
+                echo "  DELETING ${TAG} (ID: ${ID}, ${SIZE_MB} MB)..."
+                curl -s -X DELETE -H "Authorization: token ${BUILD_TOKEN}" \
+                  "${REPO_API}/releases/${ID}"
+              fi
+            done
+          else
+            echo "  Nothing to clean up"
+          fi
+
+          echo ""
+          echo "=== Done ==="

.gitea/workflows/release.yml

@@ -1,14 +1,7 @@
 name: Release
 
 on:
-  push:
+  workflow_dispatch:
-    branches: [main]
-    paths:
-      - 'src/**'
-      - 'src-tauri/**'
-      - 'package.json'
-      - 'vite.config.ts'
-      - 'index.html'
 
 jobs:
   test:
@@ -43,9 +36,6 @@ jobs:
     name: Bump version and tag
     needs: test
     runs-on: ubuntu-latest
-    outputs:
-      new_version: ${{ steps.bump.outputs.new_version }}
-      tag: ${{ steps.bump.outputs.tag }}
     steps:
       - uses: actions/checkout@v4
         with:
@@ -57,7 +47,6 @@ jobs:
           git config user.email "actions@gitea.local"
 
       - name: Bump patch version
-        id: bump
         run: |
           CURRENT=$(grep '"version"' package.json | head -1 | sed 's/.*"version": *"\([^"]*\)".*/\1/')
           echo "Current version: ${CURRENT}"
@@ -75,35 +64,34 @@ jobs:
           sed -i "s/__version__ = \"${CURRENT}\"/__version__ = \"${NEW_VERSION}\"/" version.py
           sed -i "s/__version_info__ = .*/__version_info__ = (${MAJOR}, ${MINOR}, ${NEW_PATCH})/" version.py
 
-          echo "new_version=${NEW_VERSION}" >> $GITHUB_OUTPUT
+          # Write to env file instead of step outputs (avoids act runner bug)
-          echo "tag=v${NEW_VERSION}" >> $GITHUB_OUTPUT
+          echo "NEW_VERSION=${NEW_VERSION}" >> $GITHUB_ENV
+          echo "RELEASE_TAG=v${NEW_VERSION}" >> $GITHUB_ENV
 
       - name: Commit and tag
         env:
           BUILD_TOKEN: ${{ secrets.BUILD_TOKEN }}
         run: |
-          NEW_VERSION="${{ steps.bump.outputs.new_version }}"
           git add package.json src-tauri/tauri.conf.json src-tauri/Cargo.toml version.py
           git commit -m "chore: bump version to ${NEW_VERSION} [skip ci]"
-          git tag "v${NEW_VERSION}"
+          git tag "${RELEASE_TAG}"
 
           REMOTE_URL=$(git remote get-url origin | sed "s|://|://gitea-actions:${BUILD_TOKEN}@|")
           git pull --rebase "${REMOTE_URL}" main || true
           git push "${REMOTE_URL}" HEAD:main
-          git push "${REMOTE_URL}" "v${NEW_VERSION}"
+          git push "${REMOTE_URL}" "${RELEASE_TAG}"
 
       - name: Create Gitea release
         env:
           BUILD_TOKEN: ${{ secrets.BUILD_TOKEN }}
         run: |
           REPO_API="${GITHUB_SERVER_URL}/api/v1/repos/${GITHUB_REPOSITORY}"
-          TAG="${{ steps.bump.outputs.tag }}"
+          RELEASE_NAME="Local Transcription ${RELEASE_TAG}"
-          RELEASE_NAME="Local Transcription ${TAG}"
 
           curl -s -X POST \
             -H "Authorization: token ${BUILD_TOKEN}" \
             -H "Content-Type: application/json" \
-            -d "{\"tag_name\": \"${TAG}\", \"name\": \"${RELEASE_NAME}\", \"body\": \"Automated build.\", \"draft\": false, \"prerelease\": false}" \
+            -d "{\"tag_name\": \"${RELEASE_TAG}\", \"name\": \"${RELEASE_NAME}\", \"body\": \"Automated build.\", \"draft\": false, \"prerelease\": false}" \
             "${REPO_API}/releases"
           echo "Created release: ${RELEASE_NAME}"
 
@@ -112,54 +100,14 @@ jobs:
           BUILD_TOKEN: ${{ secrets.BUILD_TOKEN }}
         run: |
           REPO_API="${GITHUB_SERVER_URL}/api/v1/repos/${GITHUB_REPOSITORY}"
-          TAG="${{ steps.bump.outputs.tag }}"
 
           for workflow in build-app-linux.yml build-app-windows.yml build-app-macos.yml; do
-            echo "Dispatching ${workflow} for ${TAG}..."
+            echo "Dispatching ${workflow} for ${RELEASE_TAG}..."
-            HTTP_CODE=$(curl -s -o /dev/null -w "%{http_code}" -X POST \
+            HTTP_CODE=$(curl -s -w "%{http_code}" -o /tmp/dispatch_resp.txt -X POST \
               -H "Authorization: token ${BUILD_TOKEN}" \
               -H "Content-Type: application/json" \
-              -d "{\"ref\": \"main\", \"inputs\": {\"tag\": \"${TAG}\"}}" \
+              -d "{\"ref\": \"main\", \"inputs\": {\"tag\": \"${RELEASE_TAG}\"}}" \
               "${REPO_API}/actions/workflows/${workflow}/dispatches")
             echo "  -> HTTP ${HTTP_CODE}"
+            if [ "$HTTP_CODE" != "204" ]; then cat /tmp/dispatch_resp.txt; echo ""; fi
           done
-
-      - name: Clean up old app releases
-        env:
-          BUILD_TOKEN: ${{ secrets.BUILD_TOKEN }}
-        run: |
-          REPO_API="${GITHUB_SERVER_URL}/api/v1/repos/${GITHUB_REPOSITORY}"
-          KEEP=3
-          PROTECT_TAG="v1.4.0"
-
-          echo "Cleaning up old app releases (keeping latest ${KEEP} + ${PROTECT_TAG})..."
-
-          # Get all app releases (v* tags, not sidecar-v*)
-          RELEASES=$(curl -s -H "Authorization: token ${BUILD_TOKEN}" \
-            "${REPO_API}/releases?limit=50" | jq -c '[.[] | select(.tag_name | startswith("v")) | select(.tag_name | startswith("sidecar") | not)]')
-
-          TOTAL=$(echo "$RELEASES" | jq 'length')
-          echo "Found ${TOTAL} app releases"
-
-          if [ "$TOTAL" -le "$KEEP" ]; then
-            echo "Nothing to clean up"
-            exit 0
-          fi
-
-          # Skip the newest KEEP releases, delete the rest (except protected)
-          echo "$RELEASES" | jq -c ".[$KEEP:][]" | while read -r release; do
-            ID=$(echo "$release" | jq -r '.id')
-            TAG=$(echo "$release" | jq -r '.tag_name')
-
-            if [ "$TAG" = "$PROTECT_TAG" ]; then
-              echo "  Protecting ${TAG}"
-              continue
-            fi
-
-            echo "  Deleting release ${TAG} (ID: ${ID})..."
-            curl -s -X DELETE -H "Authorization: token ${BUILD_TOKEN}" \
-              "${REPO_API}/releases/${ID}"
-            # Keep the git tag -- only delete the release (assets).
-            # Deleting tags breaks builds that haven't checked out yet.
-          done
-          echo "Cleanup complete"

.gitea/workflows/sidecar-release.yml

@@ -1,14 +1,6 @@
 name: Sidecar Release
 
 on:
-  push:
-    branches: [main]
-    paths:
-      - 'client/**'
-      - 'server/**'
-      - 'backend/**'
-      - 'pyproject.toml'
-      - 'local-transcription-headless.spec'
   workflow_dispatch:
 
 jobs:
@@ -35,40 +27,17 @@ jobs:
     needs: test
     if: "!contains(github.event.head_commit.message, '[skip ci]')"
     runs-on: ubuntu-latest
-    outputs:
-      version: ${{ steps.bump.outputs.version }}
-      tag: ${{ steps.bump.outputs.tag }}
-      has_changes: ${{ steps.check_changes.outputs.has_changes }}
     steps:
       - uses: actions/checkout@v4
         with:
           fetch-depth: 2
 
-      - name: Check for backend changes
-        id: check_changes
-        run: |
-          if [ "${{ github.event_name }}" = "workflow_dispatch" ]; then
-            echo "has_changes=true" >> $GITHUB_OUTPUT
-            exit 0
-          fi
-          CHANGED=$(git diff --name-only HEAD~1 HEAD -- client/ server/ backend/ pyproject.toml local-transcription-headless.spec 2>/dev/null || echo "")
-          if [ -n "$CHANGED" ]; then
-            echo "has_changes=true" >> $GITHUB_OUTPUT
-            echo "Backend changes detected: $CHANGED"
-          else
-            echo "has_changes=false" >> $GITHUB_OUTPUT
-            echo "No backend changes detected, skipping sidecar build"
-          fi
-
       - name: Configure git
-        if: steps.check_changes.outputs.has_changes == 'true'
         run: |
           git config user.name "Gitea Actions"
           git config user.email "actions@gitea.local"
 
       - name: Bump sidecar patch version
-        if: steps.check_changes.outputs.has_changes == 'true'
-        id: bump
         run: |
           CURRENT=$(grep '^version = ' pyproject.toml | head -1 | sed 's/version = "\(.*\)"/\1/')
           echo "Current sidecar version: ${CURRENT}"
@@ -82,91 +51,52 @@ jobs:
 
           sed -i "s/^version = \"${CURRENT}\"/version = \"${NEW_VERSION}\"/" pyproject.toml
 
-          echo "version=${NEW_VERSION}" >> $GITHUB_OUTPUT
+          # Write to env file instead of step outputs (avoids act runner bug)
-          echo "tag=sidecar-v${NEW_VERSION}" >> $GITHUB_OUTPUT
+          echo "NEW_VERSION=${NEW_VERSION}" >> $GITHUB_ENV
+          echo "RELEASE_TAG=sidecar-v${NEW_VERSION}" >> $GITHUB_ENV
 
       - name: Commit and tag
-        if: steps.check_changes.outputs.has_changes == 'true'
         env:
           BUILD_TOKEN: ${{ secrets.BUILD_TOKEN }}
         run: |
-          NEW_VERSION="${{ steps.bump.outputs.version }}"
-          TAG="${{ steps.bump.outputs.tag }}"
           git add pyproject.toml
           git commit -m "chore: bump sidecar version to ${NEW_VERSION} [skip ci]"
-          git tag "${TAG}"
+          git tag "${RELEASE_TAG}"
 
           REMOTE_URL=$(git remote get-url origin | sed "s|://|://gitea-actions:${BUILD_TOKEN}@|")
           git pull --rebase "${REMOTE_URL}" main || true
           git push "${REMOTE_URL}" HEAD:main
-          git push "${REMOTE_URL}" "${TAG}"
+          git push "${REMOTE_URL}" "${RELEASE_TAG}"
 
       - name: Create Gitea release
-        if: steps.check_changes.outputs.has_changes == 'true'
         env:
           BUILD_TOKEN: ${{ secrets.BUILD_TOKEN }}
         run: |
           REPO_API="${GITHUB_SERVER_URL}/api/v1/repos/${GITHUB_REPOSITORY}"
-          TAG="${{ steps.bump.outputs.tag }}"
+          RELEASE_NAME="Sidecar v${NEW_VERSION}"
-          VERSION="${{ steps.bump.outputs.version }}"
-          RELEASE_NAME="Sidecar v${VERSION}"
 
           curl -s -X POST \
             -H "Authorization: token ${BUILD_TOKEN}" \
             -H "Content-Type: application/json" \
-            -d "{\"tag_name\": \"${TAG}\", \"name\": \"${RELEASE_NAME}\", \"body\": \"Automated sidecar build.\", \"draft\": false, \"prerelease\": false}" \
+            -d "{\"tag_name\": \"${RELEASE_TAG}\", \"name\": \"${RELEASE_NAME}\", \"body\": \"Automated sidecar build.\", \"draft\": false, \"prerelease\": false}" \
             "${REPO_API}/releases"
           echo "Created release: ${RELEASE_NAME}"
 
       - name: Trigger per-OS sidecar builds
-        if: steps.check_changes.outputs.has_changes == 'true'
         env:
           BUILD_TOKEN: ${{ secrets.BUILD_TOKEN }}
         run: |
           REPO_API="${GITHUB_SERVER_URL}/api/v1/repos/${GITHUB_REPOSITORY}"
-          TAG="${{ steps.bump.outputs.tag }}"
 
-          for workflow in build-sidecar-linux.yml build-sidecar-windows.yml build-sidecar-macos.yml; do
+          for workflow in build-sidecar-linux.yml build-sidecar-windows.yml build-sidecar-macos.yml build-sidecar-cloud.yml; do
-            echo "Dispatching ${workflow} for ${TAG}..."
+            echo "Dispatching ${workflow} for ${RELEASE_TAG}..."
             HTTP_CODE=$(curl -s -o /dev/null -w "%{http_code}" -X POST \
               -H "Authorization: token ${BUILD_TOKEN}" \
               -H "Content-Type: application/json" \
-              -d "{\"ref\": \"main\", \"inputs\": {\"tag\": \"${TAG}\"}}" \
+              -d "{\"ref\": \"main\", \"inputs\": {\"tag\": \"${RELEASE_TAG}\"}}" \
               "${REPO_API}/actions/workflows/${workflow}/dispatches")
             echo "  -> HTTP ${HTTP_CODE}"
           done
 
-      - name: Clean up old sidecar releases
+      # NOTE: Automatic cleanup disabled -- it races with async builds.
-        if: steps.check_changes.outputs.has_changes == 'true'
+      # Clean up old releases manually from the Gitea UI when needed.
-        env:
-          BUILD_TOKEN: ${{ secrets.BUILD_TOKEN }}
-        run: |
-          REPO_API="${GITHUB_SERVER_URL}/api/v1/repos/${GITHUB_REPOSITORY}"
-          KEEP=2
-
-          echo "Cleaning up old sidecar releases (keeping latest ${KEEP})..."
-
-          # Get all sidecar releases (sidecar-v* tags)
-          RELEASES=$(curl -s -H "Authorization: token ${BUILD_TOKEN}" \
-            "${REPO_API}/releases?limit=50" | jq -c '[.[] | select(.tag_name | startswith("sidecar-v"))]')
-
-          TOTAL=$(echo "$RELEASES" | jq 'length')
-          echo "Found ${TOTAL} sidecar releases"
-
-          if [ "$TOTAL" -le "$KEEP" ]; then
-            echo "Nothing to clean up"
-            exit 0
-          fi
-
-          # Skip the newest KEEP releases, delete the rest
-          echo "$RELEASES" | jq -c ".[$KEEP:][]" | while read -r release; do
-            ID=$(echo "$release" | jq -r '.id')
-            TAG=$(echo "$release" | jq -r '.tag_name')
-
-            echo "  Deleting sidecar release ${TAG} (ID: ${ID})..."
-            curl -s -X DELETE -H "Authorization: token ${BUILD_TOKEN}" \
-              "${REPO_API}/releases/${ID}"
-            # Keep the git tag -- only delete the release (assets).
-            # Deleting tags breaks builds that haven't checked out yet.
-          done
-          echo "Cleanup complete"

CLAUDE.md

@@ -11,9 +11,11 @@ Local Transcription is a cross-platform desktop application for real-time speech
 **Key Features:**
 - Cross-platform desktop app (Windows, macOS, Linux) via Tauri v2 + Svelte 5
 - Headless Python backend with FastAPI control API
-- Dual transcription modes: local Whisper or cloud Deepgram (managed/BYOK)
+- Cloud-first: defaults to Deepgram (BYOK) transcription; local Whisper also supported
+- Settings UI hides local-only options (model, VAD, timing) when in cloud mode
+- Start button gated on API key / login — shows guidance if not configured
+- Shared Captions: create rooms, share via codes, join with one click (hosted at caption.shadowdao.com)
 - Built-in web server for OBS browser source at `http://localhost:8080`
-- Optional multi-user sync via Node.js server
 - CUDA, MPS (Apple Silicon), and CPU support
 - Auto-updates, custom fonts, configurable colors
 
@@ -273,9 +275,29 @@ All per-OS build workflows can be re-run independently via `workflow_dispatch` w
 - `Info.plist` must include `NSMicrophoneUsageDescription` for mic access
 - No CUDA builds — CPU/MPS only
 
+## Code Signing
+
+Code signing is configured for Windows and macOS to eliminate install warnings (SmartScreen / Gatekeeper). See [SIGNING.md](SIGNING.md) for full setup details.
+
+**Status (as of 2026-04-10):** CI workflow changes are committed. Waiting on identity verification for both platforms before secrets can be configured.
+
+**How it works:**
+- macOS: Tauri auto-signs when `APPLE_CERTIFICATE` and related env vars are set in CI. Notarization uses App Store Connect API key.
+- Windows: Azure Artifact Signing via `signtool.exe` + dlib. CI workflow injects `signCommand` into `tauri.conf.json` at build time when `AZURE_CLIENT_ID` is set.
+- Both are no-ops when secrets aren't configured — unsigned builds work as before.
+
+**Key files:**
+- `src-tauri/Entitlements.plist` — macOS hardened runtime entitlements (mic, network)
+- `src-tauri/Info.plist` — macOS microphone usage description
+- `.gitea/workflows/build-app-macos.yml` — Apple signing + notarization
+- `.gitea/workflows/build-app-windows.yml` — Azure Artifact Signing
+
+**Secrets required (12 total):** See [SIGNING.md](SIGNING.md) for the full list — 6 Apple secrets, 6 Azure secrets.
+
 ## Related Documentation
 
 - [README.md](README.md) — User-facing documentation
 - [BUILD.md](BUILD.md) — Detailed build instructions
 - [INSTALL.md](INSTALL.md) — Installation guide
+- [SIGNING.md](SIGNING.md) — Code signing setup guide
 - [server/nodejs/README.md](server/nodejs/README.md) — Node.js server setup

README.md

@@ -7,14 +7,14 @@ A real-time speech-to-text desktop application for streamers. Runs locally on yo
 ## Features
 
 - **Real-Time Transcription**: Live speech-to-text using Whisper models with minimal latency
+- **Cloud-First**: Defaults to Deepgram cloud transcription — get started with just an API key
 - **Cross-Platform**: Native desktop app for Windows, macOS, and Linux via [Tauri](https://tauri.app/)
-- **Dual Transcription Modes**: Local (Whisper) or cloud (Deepgram) with managed billing or BYOK
+- **Dual Transcription Modes**: Cloud (Deepgram) or local (Whisper) with automatic GPU/CPU detection
-- **CPU & GPU Support**: Automatic detection of CUDA (NVIDIA), MPS (Apple Silicon), or CPU fallback
+- **Shared Captions**: Create a room and share a code so others can join — no server setup needed
-- **Advanced Voice Detection**: Dual-layer VAD (WebRTC + Silero) for accurate speech detection
 - **OBS Integration**: Built-in web server for browser source capture at `http://localhost:8080`
-- **Multi-User Sync**: Optional Node.js server to sync transcriptions across multiple users
 - **Custom Fonts**: Support for system fonts, web-safe fonts, Google Fonts, and custom font files
 - **Customizable Colors**: User-configurable colors for name, text, and background
+- **Advanced Voice Detection**: Dual-layer VAD (WebRTC + Silero) for accurate speech detection
 - **Noise Suppression**: Built-in audio preprocessing to reduce background noise
 - **Auto-Updates**: Automatic update checking with release notes display
 
@@ -87,27 +87,30 @@ For detailed build instructions, see [BUILD.md](BUILD.md).
 
 ## Usage
 
-### Standalone Mode
+### Quick Setup (Cloud — Recommended)
 
 1. Launch the application
-2. Select your microphone from the audio device dropdown
+2. Open **Settings** — the transcription mode defaults to **Cloud (Deepgram)**
-3. Choose a Whisper model (smaller = faster, larger = more accurate):
+3. Get a free API key at [console.deepgram.com](https://console.deepgram.com) and paste it in Settings
+4. Select your microphone from the audio device dropdown
+5. Click **Start Transcription**
+6. Transcriptions appear in the main window and at `http://localhost:8080`
+
+> The Start button is disabled until an API key is entered. Local-only settings (model, VAD, timing) are hidden in cloud mode to keep things simple.
+
+### Local Mode (Whisper)
+
+For offline/on-device transcription, switch to **Local (Whisper)** in Settings:
+
+1. Choose a Whisper model (smaller = faster, larger = more accurate):
    - `tiny.en` / `tiny` — Fastest, good for quick captions
    - `base.en` / `base` — Balanced speed and accuracy
    - `small.en` / `small` — Better accuracy
    - `medium.en` / `medium` — High accuracy
    - `large-v3` — Best accuracy (requires more resources)
-4. Click **Start** to begin transcription
+2. Select compute device (Auto/CUDA/CPU) and compute type
-5. Transcriptions appear in the main window and at `http://localhost:8080`
+3. Tune VAD sensitivity and timing settings as needed
-
+4. Click **Start Transcription**
-### Remote Transcription (Deepgram)
-
-Instead of local Whisper models, you can use cloud-based transcription:
-
-- **Managed mode**: Sign up via the transcription proxy for metered billing
-- **BYOK mode**: Bring your own Deepgram API key for direct access
-
-Configure in Settings > Remote Transcription.
 
 ### OBS Browser Source Setup
 
@@ -117,19 +120,43 @@ Configure in Settings > Remote Transcription.
 4. Set dimensions (e.g., 1920x300)
 5. Check "Shutdown source when not visible" for performance
 
-### Multi-User Mode (Optional)
+### Shared Captions (Multi-User)
 
-For syncing transcriptions across multiple users (e.g., multi-host streams or translation teams):
+Share live captions across multiple users using the hosted service at `https://caption.shadowdao.com/` — no server setup required.
 
-1. Deploy the Node.js server (see [server/nodejs/README.md](server/nodejs/README.md))
+#### Creating a Room
-2. In the app settings, enable **Server Sync**
+
-3. Enter the server URL (e.g., `http://your-server:3000/api/send`)
+1. Open **Settings** and enable **Shared Captions**
-4. Set a room name and passphrase (shared with other users)
+2. Click **Create Room** — this generates a room name and passphrase automatically
-5. In OBS, use the server's display URL with your room name:
+3. A **share code** is generated and copied to your clipboard
+4. Send the share code to anyone who should join
+
+#### Joining a Room
+
+1. Open **Settings** and enable **Shared Captions**
+2. Paste the share code you received into the **"Paste share code to join"** field
+3. Click **Join** — the server URL, room, and passphrase are auto-filled
+4. Click **Save**
+
+#### Sharing an Existing Room
+
+If you already have a room configured and want to invite others:
+
+1. Open **Settings** and scroll to **Shared Captions**
+2. Click **Share Current Room** — generates a share code from your current config and copies it to the clipboard
+3. Send the code to others
+
+#### OBS Display for Shared Rooms
+
+In OBS, add a Browser source pointing to the server's display URL:
 ```
-   http://your-server:3000/display?room=YOURROOM&timestamps=true&maxlines=50
+https://caption.shadowdao.com/display?room=YOURROOM&timestamps=true&maxlines=50
 ```
 
+#### Self-Hosting
+
+You can also self-host the sync server. See [server/nodejs/README.md](server/nodejs/README.md) for setup instructions, then enter your own server URL in the Shared Captions settings.
+
 ## Configuration
 
 Settings are stored at `~/.local-transcription/config.yaml` and can be modified through the GUI settings panel or the REST API.
@@ -144,7 +171,7 @@ Settings are stored at `~/.local-transcription/config.yaml` and can be modified
 | `transcription.silero_sensitivity` | VAD sensitivity (0-1, lower = more sensitive) | `0.4` |
 | `transcription.post_speech_silence_duration` | Silence before finalizing (seconds) | `0.3` |
 | `transcription.continuous_mode` | Fast speaker mode for quick talkers | `false` |
-| `remote.mode` | Transcription mode (local/managed/byok) | `local` |
+| `remote.mode` | Transcription mode (local/managed/byok) | `byok` |
 | `display.show_timestamps` | Show timestamps with transcriptions | `true` |
 | `display.fade_after_seconds` | Fade out time (0 = never) | `10` |
 | `display.font_source` | Font type (System Font/Web-Safe/Google Font/Custom File) | `System Font` |
@@ -267,6 +294,15 @@ Both workflows require a `BUILD_TOKEN` secret in the repo settings (Gitea API to
 
 ## Troubleshooting
 
+### macOS: "App is damaged and can't be opened"
+macOS Gatekeeper blocks unsigned applications. Since the app is not yet signed with an Apple Developer certificate, you need to remove the quarantine flag before opening:
+
+```bash
+xattr -cr "/Applications/Local Transcription.app"
+```
+
+Then open the app normally. You only need to do this once after downloading.
+
 ### Model Loading Issues
 - Models download automatically on first use to `~/.cache/huggingface/`
 - First run requires internet connection

SIGNING.md

@@ -0,0 +1,136 @@
+# Code Signing Setup
+
+This document explains how to configure code signing for Local Transcription so that Windows and macOS installers are trusted by the operating system.
+
+## Overview
+
+Without code signing:
+- **Windows**: SmartScreen shows "Windows protected your PC" warnings
+- **macOS**: Gatekeeper blocks the app — "app can't be opened because it is from an unidentified developer"
+
+The CI/CD workflows are configured to sign automatically when the required secrets are present. Without secrets, builds still work — they just produce unsigned installers.
+
+---
+
+## Windows — Azure Artifact Signing
+
+**Cost**: ~$9.99/month (up to 5,000 signatures)
+
+### 1. Create an Azure Account
+
+Sign up at https://azure.microsoft.com if you don't already have one.
+
+### 2. Set Up Artifact Signing
+
+1. In the Azure Portal, search for **Artifact Signing**
+2. Create a new **Artifact Signing Account**
+   - Choose a region (e.g., West US 2) — note this for the endpoint URL
+   - The endpoint will be like `https://wus2.codesigning.azure.net/`
+3. Complete **Identity Verification** (required before you can create certificate profiles)
+4. Create a **Certificate Profile** with type "Public Trust" for code signing
+
+### 3. Create an App Registration (Service Principal)
+
+This allows CI to authenticate to Azure:
+
+1. Go to **Azure Active Directory** > **App registrations** > **New registration**
+2. Name it (e.g., `local-transcription-signing`)
+3. After creation, note the **Application (client) ID** and **Directory (tenant) ID**
+4. Go to **Certificates & secrets** > **New client secret** — note the secret value
+5. Grant the app registration the **Artifact Signing Certificate Profile Signer** role on your Artifact Signing Account
+
+### 4. Add Gitea Secrets
+
+In your Gitea repository, go to **Settings** > **Actions** > **Secrets** and add:
+
+| Secret Name | Value |
+|-------------|-------|
+| `AZURE_CLIENT_ID` | App registration Application (client) ID |
+| `AZURE_CLIENT_SECRET` | App registration client secret value |
+| `AZURE_TENANT_ID` | Directory (tenant) ID |
+| `AZURE_SIGNING_ENDPOINT` | Artifact Signing endpoint URL (e.g., `https://wus2.codesigning.azure.net/`) |
+| `AZURE_SIGNING_ACCOUNT` | Artifact Signing account name |
+| `AZURE_CERT_PROFILE` | Certificate profile name |
+
+---
+
+## macOS — Apple Developer Code Signing + Notarization
+
+**Cost**: $99/year (Apple Developer Program)
+
+### 1. Enroll in the Apple Developer Program
+
+Sign up at https://developer.apple.com/programs/
+
+### 2. Create a Developer ID Certificate
+
+1. Open **Xcode** > **Settings** > **Accounts** > select your team > **Manage Certificates**
+2. Click **+** > **Developer ID Application**
+3. Or create via the Apple Developer portal: **Certificates, Identifiers & Profiles** > **Certificates** > **+** > **Developer ID Application**
+
+### 3. Export the Certificate as .p12
+
+1. Open **Keychain Access**
+2. Find your **Developer ID Application** certificate
+3. Right-click > **Export** > save as `.p12` with a password
+4. Base64-encode it:
+   ```bash
+   base64 -i certificate.p12 | tr -d '\n'
+   ```
+
+### 4. Create an App Store Connect API Key
+
+This is used for notarization (submitting the app to Apple for verification):
+
+1. Go to https://appstoreconnect.apple.com/access/integrations/api
+2. Click **Generate API Key**
+3. Give it a name and **Developer** role (minimum)
+4. Download the `.p8` private key file (you can only download it once)
+5. Note the **Key ID** and **Issuer ID** shown on the page
+
+### 5. Find Your Signing Identity
+
+Your signing identity looks like:
+```
+Developer ID Application: Your Name (TEAMID)
+```
+
+You can find it by running:
+```bash
+security find-identity -v -p codesigning
+```
+
+### 6. Add Gitea Secrets
+
+| Secret Name | Value |
+|-------------|-------|
+| `APPLE_CERTIFICATE` | Base64-encoded .p12 certificate (from step 3) |
+| `APPLE_CERTIFICATE_PASSWORD` | Password used when exporting the .p12 |
+| `APPLE_SIGNING_IDENTITY` | Full identity string (e.g., `Developer ID Application: Your Name (TEAMID)`) |
+| `APPLE_API_KEY` | App Store Connect API Key ID |
+| `APPLE_API_ISSUER` | API issuer UUID |
+| `APPLE_API_KEY_CONTENT` | Full contents of the `.p8` private key file |
+
+---
+
+## Verifying Signing Works
+
+### Trigger a Build
+
+Both build workflows use `workflow_dispatch`, so you can trigger them manually in Gitea:
+
+1. Go to **Actions** > select the workflow > **Run workflow**
+2. Enter the release tag (e.g., `v2.0.15`)
+
+### Check macOS
+
+After installing the `.dmg`, the app should open without any Gatekeeper warnings. You can also verify from the command line:
+
+```bash
+codesign -dv --verbose=4 /Applications/Local\ Transcription.app
+spctl --assess --type execute /Applications/Local\ Transcription.app
+```
+
+### Check Windows
+
+After running the `.msi` or `-setup.exe`, there should be no SmartScreen warning. The installer properties should show your organization name as the publisher.

backend/api_server.py

@@ -73,8 +73,15 @@ class APIServer:
         original_state_cb = self.controller.on_state_changed
 
         def on_state_changed(state: str, message: str):
+            # Isolate the upstream callback so a failure there (e.g. a
+            # broken stdout pipe in main_headless) cannot propagate into
+            # _set_state and tear down engine init / reload_engine /
+            # apply_settings request handling.
             if original_state_cb:
+                try:
                     original_state_cb(state, message)
+                except Exception:
+                    pass
             self._broadcast_control({"type": "state_changed", "state": state, "message": message})
 
         self.controller.on_state_changed = on_state_changed
@@ -151,14 +158,24 @@ class APIServer:
 
         @app.post("/api/start")
         async def start_transcription():
-            success, message = ctrl.start_transcription()
+            import asyncio
+            # Run in thread pool to avoid blocking the event loop
+            # (start_recording can block up to 15s waiting for Deepgram WS)
+            loop = asyncio.get_event_loop()
+            success, message = await loop.run_in_executor(
+                None, ctrl.start_transcription
+            )
             if not success:
                 raise HTTPException(status_code=400, detail=message)
             return {"status": "ok", "message": message}
 
         @app.post("/api/stop")
         async def stop_transcription():
-            success, message = ctrl.stop_transcription()
+            import asyncio
+            loop = asyncio.get_event_loop()
+            success, message = await loop.run_in_executor(
+                None, ctrl.stop_transcription
+            )
             if not success:
                 raise HTTPException(status_code=400, detail=message)
             return {"status": "ok", "message": message}
@@ -202,7 +219,11 @@ class APIServer:
 
         @app.put("/api/config")
         async def update_config(update: ConfigUpdate):
-            engine_reloaded, message = ctrl.apply_settings(update.settings)
+            import asyncio
+            loop = asyncio.get_event_loop()
+            engine_reloaded, message = await loop.run_in_executor(
+                None, ctrl.apply_settings, update.settings
+            )
             return {
                 "status": "ok",
                 "message": message,
@@ -223,7 +244,11 @@ class APIServer:
 
         @app.post("/api/reload-engine")
         async def reload_engine():
-            success, message = ctrl.reload_engine()
+            import asyncio
+            loop = asyncio.get_event_loop()
+            success, message = await loop.run_in_executor(
+                None, ctrl.reload_engine
+            )
             if not success:
                 raise HTTPException(status_code=500, detail=message)
             return {"status": "ok", "message": message}
@@ -255,6 +280,7 @@ class APIServer:
                     data = resp.json()
                     ctrl.config.set('remote.auth_token', data.get('token', ''))
                     ctrl.config.set('remote.server_url', req.server_url)
+                    ctrl.config.set('remote.email', req.email)
                     return {"status": "ok", "token": data.get('token', '')}
                 else:
                     raise HTTPException(status_code=resp.status_code, detail=resp.text)

backend/app_controller.py

@@ -18,13 +18,18 @@ import sys
 sys.path.insert(0, str(Path(__file__).resolve().parent.parent))
 
 from client.config import Config
-from client.device_utils import DeviceManager
+from client.models import TranscriptionResult
-from client.transcription_engine_realtime import RealtimeTranscriptionEngine, TranscriptionResult
 from client.deepgram_transcription import DeepgramTranscriptionEngine
 from client.server_sync import ServerSyncClient
 from server.web_display import TranscriptionWebServer
 from version import __version__
 
+# Heavy imports (torch, RealtimeSTT, faster-whisper) are deferred so
+# the cloud-only sidecar build can exclude them entirely.
+# Imported lazily in _initialize_engine() when remote.mode == "local".
+RealtimeTranscriptionEngine = None
+DeviceManager = None
+
 
 class AppState:
     """Enum-like class for application states."""
@@ -89,7 +94,24 @@ class AppController:
 
     def __init__(self, config: Optional[Config] = None):
         self.config = config or Config()
-        self.device_manager = DeviceManager()
+
+        # DeviceManager is only needed for local Whisper mode.
+        # Lazy-import to keep the cloud-only sidecar lightweight.
+        global DeviceManager
+        if DeviceManager is None:
+            try:
+                from client.device_utils import DeviceManager as _DM
+                DeviceManager = _DM
+            except ImportError:
+                DeviceManager = None
+
+        self.device_manager = DeviceManager() if DeviceManager else None
+        self.is_cloud_only = DeviceManager is None
+
+        # If this is the cloud-only sidecar and mode is still "local",
+        # auto-switch to "byok" so the engine doesn't try to load Whisper.
+        if self.is_cloud_only and self.config.get('remote.mode', 'local') == 'local':
+            self.config.set('remote.mode', 'byok')
 
         # State
         self._state = AppState.INITIALIZING
@@ -243,21 +265,17 @@ class AppController:
 
     def _initialize_engine(self):
         """Initialize the transcription engine in a background thread."""
-        device_config = self.config.get('transcription.device', 'auto')
-        self.device_manager.set_device(device_config)
-
         audio_device_str = self.config.get('audio.input_device', 'default')
         audio_device = None if audio_device_str == 'default' else int(audio_device_str)
 
         model = self.config.get('transcription.model', 'base.en')
         language = self.config.get('transcription.language', 'en')
-        device = self.device_manager.get_device_for_whisper()
+        device_config = self.config.get('transcription.device', 'auto')
         compute_type = self.config.get('transcription.compute_type', 'default')
 
         self.current_model_size = model
         self.current_device_config = device_config
 
-        user_name = self.config.get('user.name', 'User')
         continuous_mode = self.config.get('transcription.continuous_mode', False)
 
         if continuous_mode:
@@ -274,7 +292,6 @@ class AppController:
         if remote_mode in ('managed', 'byok'):
             self.transcription_engine = DeepgramTranscriptionEngine(
                 config=self.config,
-                user_name=user_name,
                 input_device_index=audio_device,
             )
             self.transcription_engine.set_callbacks(
@@ -284,6 +301,27 @@ class AppController:
             self.transcription_engine.set_error_callback(self._on_remote_error)
             self.transcription_engine.set_credits_low_callback(self._on_credits_low)
         else:
+            # Lazy-import heavy local transcription dependencies
+            global RealtimeTranscriptionEngine
+            if RealtimeTranscriptionEngine is None:
+                try:
+                    from client.transcription_engine_realtime import RealtimeTranscriptionEngine as _RTE
+                    RealtimeTranscriptionEngine = _RTE
+                except ImportError:
+                    # Cloud-only sidecar -- local engine not available
+                    self._set_state(
+                        AppState.ERROR,
+                        "Local transcription not available in this build. "
+                        "Please switch to Cloud (Deepgram) mode in Settings."
+                    )
+                    return
+
+            if self.device_manager:
+                self.device_manager.set_device(device_config)
+                device = self.device_manager.get_device_for_whisper()
+            else:
+                device = "cpu"
+
             self.transcription_engine = RealtimeTranscriptionEngine(
                 model=model,
                 device=device,
@@ -303,7 +341,7 @@ class AppController:
                 initial_prompt=self.config.get('transcription.initial_prompt', ''),
                 no_log_file=self.config.get('transcription.no_log_file', True),
                 input_device_index=audio_device,
-                user_name=user_name,
+                app_config=self.config,
             )
             self.transcription_engine.set_callbacks(
                 realtime_callback=self._on_realtime_transcription,
@@ -332,6 +370,14 @@ class AppController:
                 device_display = "Unknown"
 
             self._set_state(AppState.READY, f"Ready | Device: {device_display}")
+        else:
+            # Cloud sidecar with no API key -- show helpful setup message
+            # instead of a scary error. The user needs to enter their key.
+            if self.is_cloud_only:
+                self._set_state(
+                    AppState.READY,
+                    "Setup needed: Open Settings > Remote Transcription > enter your Deepgram API key"
+                )
             else:
                 self._set_state(AppState.ERROR, message)
 
@@ -348,7 +394,14 @@ class AppController:
         try:
             success = self.transcription_engine.start_recording()
             if not success:
-                return False, "Failed to start recording"
+                import logging
+                # Check if there's a recent error in the logger
+                err_detail = getattr(self.transcription_engine, '_last_error', '')
+                msg = f"Failed to start recording"
+                if err_detail:
+                    msg += f": {err_detail}"
+                print(f"ERROR: {msg}")
+                return False, msg
 
             # Start server sync if enabled
             if self.config.get('server_sync.enabled', False):
@@ -553,8 +606,17 @@ class AppController:
         Returns (engine_reload_needed, message).
         """
         if new_config:
-            for key, value in new_config.items():
+            # Flatten nested dicts into dot-notation keys so we merge
-                self.config.set(key, value)
+            # individual values instead of replacing entire sections
+            # (e.g. remote.mode instead of overwriting all of remote)
+            def _flatten(d, prefix=""):
+                for k, v in d.items():
+                    full_key = f"{prefix}{k}" if not prefix else f"{prefix}.{k}"
+                    if isinstance(v, dict):
+                        _flatten(v, full_key)
+                    else:
+                        self.config.set(full_key, v)
+            _flatten(new_config)
 
         # Update web server display settings
         if self.web_server:
@@ -577,12 +639,18 @@ class AppController:
             if self.config.get('server_sync.enabled', False):
                 self._start_server_sync()
 
-        # Check if model/device changed
+        # Check if model/device/remote mode changed -- any of these require
+        # a full engine reload since they change which engine class is used
         new_model = self.config.get('transcription.model', 'base.en')
         new_device = self.config.get('transcription.device', 'auto')
+        new_remote_mode = self.config.get('remote.mode', 'local')
+        current_remote_mode = 'local'
+        if self.transcription_engine:
+            current_remote_mode = getattr(self.transcription_engine, 'mode', 'local')
         engine_reload_needed = (
             self.current_model_size != new_model
             or self.current_device_config != new_device
+            or current_remote_mode != new_remote_mode
         )
 
         if engine_reload_needed:
@@ -596,7 +664,7 @@ class AppController:
         host = self.config.get('web_server.host', '127.0.0.1')
         port = self.actual_web_port or self.config.get('web_server.port', 8080)
 
-        device_info = self.device_manager.get_device_info()
+        device_info = self.device_manager.get_device_info() if self.device_manager else []
 
         remote_mode = self.config.get('remote.mode', 'local')
         if remote_mode in ('managed', 'byok') and self.transcription_engine:
@@ -621,6 +689,7 @@ class AppController:
             "transcription_count": len(self.transcriptions),
             "remote_mode": remote_mode,
             "server_sync_enabled": self.config.get('server_sync.enabled', False),
+            "is_cloud_only": self.is_cloud_only,
         }
 
     def get_audio_devices(self) -> list[dict]:
@@ -640,10 +709,13 @@ class AppController:
 
     def get_compute_devices(self) -> list[dict]:
         """List available compute devices."""
-        device_info = self.device_manager.get_device_info()
         devices = [{"id": "auto", "name": "Auto-detect"}]
+        if self.device_manager:
+            device_info = self.device_manager.get_device_info()
             for dev_id, dev_name in device_info:
                 devices.append({"id": dev_id, "name": dev_name})
+        else:
+            devices.append({"id": "cloud", "name": "Cloud (Deepgram)"})
         return devices
 
     # ── Update Checking ────────────────────────────────────────────

backend/main_headless.py

@@ -75,10 +75,16 @@ def main():
         # Create controller and initialize
         controller = AppController(config=config)
 
-        # Wire a state callback that prints the ready event
+        # Wire a state callback that prints state events for the parent
+        # process to read. Stdout writes can fail with EINVAL on Windows
+        # when the parent stops reading the sidecar pipe; swallow those
+        # so the engine state machine isn't taken down by a logging path.
         def on_state_changed(state, message):
             event = {"event": "state", "state": state, "message": message}
+            try:
                 print(json.dumps(event), flush=True)
+            except (OSError, ValueError):
+                pass
 
         controller.on_state_changed = on_state_changed
 

backend/tests/test_api_server.py

@@ -79,7 +79,7 @@ async def test_start_when_not_ready(api_client, controller):
 
 @pytest.mark.asyncio
 async def test_clear(api_client, controller):
-    from client.transcription_engine_realtime import TranscriptionResult
+    from client.models import TranscriptionResult
     from datetime import datetime
 
     controller.transcriptions = [

backend/tests/test_app_controller.py

@@ -72,7 +72,7 @@ def test_double_start_rejected(controller):
 
 def test_clear_transcriptions(controller):
     """clear_transcriptions should empty the list and return the count."""
-    from client.transcription_engine_realtime import TranscriptionResult
+    from client.models import TranscriptionResult
 
     controller.transcriptions = [
         TranscriptionResult(text="Hello", is_final=True, timestamp=datetime.now(), user_name="Alice"),
@@ -85,7 +85,7 @@ def test_clear_transcriptions(controller):
 
 def test_get_transcriptions_text_with_timestamps(controller):
     """get_transcriptions_text should include [HH:MM:SS] prefixes when requested."""
-    from client.transcription_engine_realtime import TranscriptionResult
+    from client.models import TranscriptionResult
 
     ts = datetime(2025, 1, 15, 10, 30, 45)
     controller.transcriptions = [
@@ -125,6 +125,8 @@ def test_apply_settings_no_reload_when_same(controller):
     # Ensure config returns the same values
     controller.config.set("transcription.model", "base.en")
     controller.config.set("transcription.device", "auto")
+    # Remote mode must also match (no engine means current mode is 'local')
+    controller.config.set("remote.mode", "local")
 
     controller.reload_engine = MagicMock(return_value=(True, "reloaded"))
 
@@ -141,7 +143,7 @@ def test_apply_settings_no_reload_when_same(controller):
 
 def test_on_final_transcription_callback_fires(controller):
     """_on_final_transcription should append and invoke on_transcription callback."""
-    from client.transcription_engine_realtime import TranscriptionResult
+    from client.models import TranscriptionResult
 
     received = []
     controller.on_transcription = lambda data: received.append(data)
@@ -166,7 +168,7 @@ def test_on_final_transcription_callback_fires(controller):
 
 def test_on_final_transcription_ignored_when_not_transcribing(controller):
     """If the controller is not in transcribing state the callback should be a no-op."""
-    from client.transcription_engine_realtime import TranscriptionResult
+    from client.models import TranscriptionResult
 
     controller.is_transcribing = False
 

client/deepgram_transcription.py

@@ -17,7 +17,7 @@ from datetime import datetime
 from queue import Queue, Empty
 from typing import Optional, Callable
 
-from client.transcription_engine_realtime import TranscriptionResult
+from client.models import TranscriptionResult
 
 logger = logging.getLogger(__name__)
 
@@ -36,18 +36,16 @@ class DeepgramTranscriptionEngine:
     #  Construction / configuration
     # ------------------------------------------------------------------ #
 
-    def __init__(self, config, user_name: str = "User", input_device_index: Optional[int] = None):
+    def __init__(self, config, input_device_index: Optional[int] = None):
         """
         Initialise the engine from a :class:`client.config.Config` object.
 
         Args:
             config: Application ``Config`` instance.
-            user_name: Display name attached to transcriptions.
             input_device_index: Index of the audio input device to use
                 (``None`` for the system default).
         """
         self.config = config
-        self.user_name = user_name
         self.input_device_index = input_device_index
 
         # Mode: 'managed' (proxy) or 'byok' (direct Deepgram)
@@ -67,7 +65,7 @@ class DeepgramTranscriptionEngine:
         # Audio parameters
         self.sample_rate: int = 16000
         self.channels: int = 1
-        self.blocksize: int = 4096
+        self.blocksize: int = 1024  # ~64ms chunks for lower latency streaming
 
         # Callbacks
         self.realtime_callback: Optional[Callable[[TranscriptionResult], None]] = None
@@ -156,17 +154,30 @@ class DeepgramTranscriptionEngine:
             return True
 
         self._stop_event.clear()
+        self._ws_connected = threading.Event()
         self._is_recording = True
 
         # Start the asyncio event-loop thread (handles WS send/receive)
         self._thread = threading.Thread(target=self._run_event_loop, daemon=True)
         self._thread.start()
 
+        # Wait for the WebSocket to connect before starting audio capture.
+        # Without this, audio chunks arrive before the WS is open -> broken pipe.
+        if not self._ws_connected.wait(timeout=15):
+            logger.error("Timed out waiting for Deepgram WebSocket connection")
+            print("ERROR: Timed out waiting for Deepgram WebSocket connection")
+            self._last_error = "Timed out connecting to Deepgram"
+            self._is_recording = False
+            self._stop_event.set()
+            return False
+
         # Start the audio capture stream
         try:
             self._start_audio_stream()
         except Exception as exc:
             logger.error("Failed to open audio stream: %s", exc)
+            print(f"ERROR: Failed to open audio stream: {exc}")
+            self._last_error = f"Audio stream error: {exc}"
             self._is_recording = False
             self._stop_event.set()
             return False
@@ -283,6 +294,11 @@ class DeepgramTranscriptionEngine:
                 if not await self._managed_handshake():
                     return
 
+            # Signal that the WebSocket is connected and ready
+            logger.info("WebSocket connected to Deepgram")
+            if hasattr(self, '_ws_connected'):
+                self._ws_connected.set()
+
             # Run send and receive concurrently
             await asyncio.gather(
                 self._send_loop(),
@@ -302,9 +318,13 @@ class DeepgramTranscriptionEngine:
     def _build_ws_url_and_headers(self):
         """Return ``(url, headers)`` depending on the current mode."""
         if self.mode == "managed":
-            # Ensure the server URL uses wss:// and append the path
+            # Convert HTTP(S) URLs to WS(S) for WebSocket connection
             url = self.server_url.rstrip("/")
-            if not url.startswith("ws://") and not url.startswith("wss://"):
+            if url.startswith("https://"):
+                url = "wss://" + url[len("https://"):]
+            elif url.startswith("http://"):
+                url = "ws://" + url[len("http://"):]
+            elif not url.startswith("ws://") and not url.startswith("wss://"):
                 url = f"wss://{url}"
             url = f"{url}/ws/transcribe"
             return url, {}
@@ -314,6 +334,8 @@ class DeepgramTranscriptionEngine:
             f"model={self.deepgram_model}"
             f"&language={self.language}"
             "&interim_results=true"
+            "&punctuate=true"
+            "&smart_format=true"
             "&encoding=linear16"
             f"&sample_rate={self.sample_rate}"
             f"&channels={self.channels}"
@@ -370,10 +392,16 @@ class DeepgramTranscriptionEngine:
 
     async def _send_loop(self):
         """Drain the audio queue and push raw PCM bytes over the WebSocket."""
+        loop = asyncio.get_event_loop()
         while not self._stop_event.is_set():
             try:
-                pcm_bytes = self._audio_queue.get(timeout=0.1)
+                # Use run_in_executor to avoid blocking the async event loop
-            except Empty:
+                # (which would stall the receive loop and delay transcriptions)
+                pcm_bytes = await asyncio.wait_for(
+                    loop.run_in_executor(None, lambda: self._audio_queue.get(timeout=0.5)),
+                    timeout=1.0,
+                )
+            except (Empty, asyncio.TimeoutError):
                 continue
 
             try:
@@ -424,7 +452,7 @@ class DeepgramTranscriptionEngine:
                     text=text,
                     is_final=is_final,
                     timestamp=datetime.now(),
-                    user_name=self.user_name,
+                    user_name=self.config.get('user.name', 'User'),
                 )
                 if is_final:
                     if self.final_callback:
@@ -475,7 +503,7 @@ class DeepgramTranscriptionEngine:
                     text=transcript,
                     is_final=is_final,
                     timestamp=datetime.now(),
-                    user_name=self.user_name,
+                    user_name=self.config.get('user.name', 'User'),
                 )
                 if is_final:
                     if self.final_callback:
@@ -506,10 +534,6 @@ class DeepgramTranscriptionEngine:
                 pass
             self._ws = None
 
-    def set_user_name(self, user_name: str):
-        """Update the user name attached to future transcriptions."""
-        self.user_name = user_name
-
     def is_recording_active(self) -> bool:
         """Return ``True`` if audio is currently being captured."""
         return self._is_recording

client/models.py

@@ -0,0 +1,29 @@
+"""Shared data models used across transcription engines."""
+
+from datetime import datetime
+
+
+class TranscriptionResult:
+    """Represents a transcription result."""
+
+    def __init__(self, text: str, is_final: bool, timestamp: datetime, user_name: str = ""):
+        """
+        Initialize transcription result.
+
+        Args:
+            text: Transcribed text
+            is_final: Whether this is a final transcription or realtime preview
+            timestamp: Timestamp of transcription
+            user_name: Name of the user/speaker
+        """
+        self.text = text.strip()
+        self.is_final = is_final
+        self.timestamp = timestamp
+        self.user_name = user_name
+
+    def __repr__(self) -> str:
+        time_str = self.timestamp.strftime("%H:%M:%S")
+        prefix = "[FINAL]" if self.is_final else "[PREVIEW]"
+        if self.user_name and self.user_name.strip():
+            return f"{prefix} [{time_str}] {self.user_name}: {self.text}"
+        return f"{prefix} [{time_str}] {self.text}"

client/transcription_engine_realtime.py

@@ -8,30 +8,8 @@ from threading import Lock
 import logging
 
 
-class TranscriptionResult:
+# Re-export TranscriptionResult from the shared models module for backward compatibility
-    """Represents a transcription result."""
+from client.models import TranscriptionResult  # noqa: F401
-
-    def __init__(self, text: str, is_final: bool, timestamp: datetime, user_name: str = ""):
-        """
-        Initialize transcription result.
-
-        Args:
-            text: Transcribed text
-            is_final: Whether this is a final transcription or realtime preview
-            timestamp: Timestamp of transcription
-            user_name: Name of the user/speaker
-        """
-        self.text = text.strip()
-        self.is_final = is_final
-        self.timestamp = timestamp
-        self.user_name = user_name
-
-    def __repr__(self) -> str:
-        time_str = self.timestamp.strftime("%H:%M:%S")
-        prefix = "[FINAL]" if self.is_final else "[PREVIEW]"
-        if self.user_name and self.user_name.strip():
-            return f"{prefix} [{time_str}] {self.user_name}: {self.text}"
-        return f"{prefix} [{time_str}] {self.text}"
 
     def to_dict(self) -> dict:
         """Convert to dictionary."""
@@ -80,8 +58,8 @@ class RealtimeTranscriptionEngine:
         no_log_file: bool = True,
         # Audio device
         input_device_index: Optional[int] = None,
-        # User name
+        # App config (for reading user.name at transcription time)
-        user_name: str = ""
+        app_config=None
     ):
         """
         Initialize RealtimeSTT transcription engine.
@@ -104,7 +82,7 @@ class RealtimeTranscriptionEngine:
             initial_prompt: Optional prompt to guide transcription
             no_log_file: Disable RealtimeSTT logging
             input_device_index: Audio input device index
-            user_name: User name for transcriptions
+            app_config: App Config object for reading user.name dynamically
         """
         self.model = model
         self.language = language
@@ -122,7 +100,7 @@ class RealtimeTranscriptionEngine:
         self.enable_realtime = enable_realtime_transcription
         self.realtime_model = realtime_model
         self.realtime_processing_pause = realtime_processing_pause
-        self.user_name = user_name
+        self.app_config = app_config
 
         # Callbacks
         self.realtime_callback: Optional[Callable[[TranscriptionResult], None]] = None
@@ -184,6 +162,11 @@ class RealtimeTranscriptionEngine:
         self.realtime_callback = realtime_callback
         self.final_callback = final_callback
 
+    def _get_user_name(self) -> str:
+        if self.app_config:
+            return self.app_config.get('user.name', '')
+        return ''
+
     def _on_realtime_transcription(self, text: str):
         """Internal callback for realtime transcriptions."""
         if self.realtime_callback and text.strip():
@@ -191,7 +174,7 @@ class RealtimeTranscriptionEngine:
                 text=text,
                 is_final=False,
                 timestamp=datetime.now(),
-                user_name=self.user_name
+                user_name=self._get_user_name()
             )
             self.realtime_callback(result)
 
@@ -202,7 +185,7 @@ class RealtimeTranscriptionEngine:
                 text=text,
                 is_final=True,
                 timestamp=datetime.now(),
-                user_name=self.user_name
+                user_name=self._get_user_name()
             )
             self.final_callback(result)
 
@@ -428,10 +411,6 @@ class RealtimeTranscriptionEngine:
         if self.is_recording:
             print("VAD settings updated. Restart transcription to apply changes.")
 
-    def set_user_name(self, user_name: str):
-        """Set the user name for transcriptions."""
-        self.user_name = user_name
-
     def __repr__(self) -> str:
         return f"RealtimeTranscriptionEngine(model={self.model}, device={self.device}, running={self.is_recording})"
 

config/default_config.yaml

@@ -42,7 +42,7 @@ transcription:
 
 server_sync:
   enabled: false
-  url: "http://localhost:3000/api/send"
+  url: ""
   room: "default"
   passphrase: ""
   # Font settings are now in the display section (shared for local and server sync)
@@ -69,9 +69,10 @@ web_server:
   host: "127.0.0.1"
 
 remote:
-  mode: local  # local | managed | byok
+  mode: byok  # local | managed | byok
-  server_url: ""  # Proxy server URL for managed mode (e.g., wss://your-proxy.com)
+  server_url: "https://transcribe.shadowdao.com"  # Proxy server URL for managed mode
   auth_token: ""  # JWT stored after login (managed mode)
+  email: ""  # Email of the logged-in managed-mode account (for UI display)
   byok_api_key: ""  # Deepgram API key for BYOK mode
   deepgram_model: nova-2  # Deepgram model to use
   language: en-US  # Language code

gui/main_window_qt.py

@@ -401,7 +401,6 @@ class MainWindow(QMainWindow):
             # Use Deepgram-based remote transcription
             self.transcription_engine = DeepgramTranscriptionEngine(
                 config=self.config,
-                user_name=user_name,
                 input_device_index=audio_device
             )
             self.transcription_engine.set_callbacks(
@@ -431,7 +430,7 @@ class MainWindow(QMainWindow):
                 initial_prompt=self.config.get('transcription.initial_prompt', ''),
                 no_log_file=self.config.get('transcription.no_log_file', True),
                 input_device_index=audio_device,
-                user_name=user_name
+                app_config=self.config
             )
 
         # Set up callbacks for transcription results

local-transcription-cloud.spec

@@ -0,0 +1,169 @@
+# -*- mode: python ; coding: utf-8 -*-
+"""PyInstaller spec file for cloud-only Local Transcription backend.
+
+This builds a lightweight sidecar (~50MB) that only supports Deepgram
+cloud transcription (managed + BYOK). No local Whisper models, no
+PyTorch, no CUDA -- just audio capture and WebSocket streaming.
+"""
+
+import sys
+import os
+
+block_cipher = None
+is_windows = sys.platform == 'win32'
+
+from PyInstaller.utils.hooks import collect_submodules, collect_data_files
+
+# Data files
+datas = [
+    ('config/default_config.yaml', 'config'),
+]
+
+# Collect sounddevice's bundled PortAudio library (_sounddevice_data)
+try:
+    import sounddevice
+    sd_path = os.path.dirname(sounddevice.__file__)
+    sd_data = os.path.join(sd_path, '_sounddevice_data')
+    if os.path.exists(sd_data):
+        datas.append((sd_data, '_sounddevice_data'))
+        print(f"  + Collected sounddevice PortAudio data from {sd_data}")
+    # Also collect the package itself
+    sd_datas = collect_data_files('sounddevice')
+    if sd_datas:
+        datas += sd_datas
+        print(f"  + Collected {len(sd_datas)} sounddevice data files")
+except ImportError:
+    print("  - Warning: sounddevice not found")
+
+# Hidden imports -- only lightweight deps needed for Deepgram streaming
+hiddenimports = [
+    'sounddevice',
+    '_sounddevice_data',
+    'numpy',
+    # FastAPI and dependencies
+    'fastapi',
+    'fastapi.routing',
+    'fastapi.responses',
+    'starlette',
+    'starlette.applications',
+    'starlette.routing',
+    'starlette.responses',
+    'starlette.websockets',
+    'starlette.middleware',
+    'starlette.middleware.cors',
+    'pydantic',
+    'pydantic.fields',
+    'pydantic.main',
+    'anyio',
+    'anyio._backends',
+    'anyio._backends._asyncio',
+    'sniffio',
+    # Uvicorn
+    'uvicorn',
+    'uvicorn.logging',
+    'uvicorn.loops',
+    'uvicorn.loops.auto',
+    'uvicorn.protocols',
+    'uvicorn.protocols.http',
+    'uvicorn.protocols.http.auto',
+    'uvicorn.protocols.http.h11_impl',
+    'uvicorn.protocols.websockets',
+    'uvicorn.protocols.websockets.auto',
+    'uvicorn.protocols.websockets.wsproto_impl',
+    'uvicorn.lifespan',
+    'uvicorn.lifespan.on',
+    'h11',
+    'websockets',
+    'websockets.legacy',
+    'websockets.legacy.server',
+    # HTTP client
+    'requests',
+    'urllib3',
+    'certifi',
+    'charset_normalizer',
+]
+
+# Collect submodules for key packages
+print("Collecting submodules for cloud backend packages...")
+for package in ['fastapi', 'starlette', 'pydantic', 'pydantic_core', 'anyio', 'uvicorn', 'websockets', 'h11']:
+    try:
+        submodules = collect_submodules(package)
+        hiddenimports += submodules
+        print(f"  + Collected {len(submodules)} submodules from {package}")
+    except Exception as e:
+        print(f"  - Warning: Could not collect {package}: {e}")
+
+# Collect data files
+for package in ['fastapi', 'starlette', 'pydantic', 'uvicorn']:
+    try:
+        data_files = collect_data_files(package)
+        if data_files:
+            datas += data_files
+    except Exception:
+        pass
+
+# Pydantic critical deps
+hiddenimports += [
+    'colorsys', 'decimal', 'json', 'ipaddress', 'pathlib', 'uuid',
+    'email.message', 'typing_extensions',
+]
+
+a = Analysis(
+    ['backend/main_headless.py'],
+    pathex=[],
+    binaries=[],
+    datas=datas,
+    hiddenimports=hiddenimports,
+    hookspath=['hooks'],
+    hooksconfig={},
+    runtime_hooks=[],
+    excludes=[
+        # Exclude all heavy ML/local transcription deps
+        'torch', 'torchaudio', 'torchvision',
+        'faster_whisper', 'ctranslate2',
+        'RealtimeSTT', 'webrtcvad', 'webrtcvad_wheels',
+        'silero_vad', 'onnxruntime',
+        'openwakeword', 'pvporcupine', 'pyaudio',
+        'noisereduce', 'scipy',
+        # Exclude GUI frameworks
+        'PySide6', 'PyQt5', 'PyQt6', 'tkinter',
+        # Exclude other unnecessary heavy packages
+        'matplotlib', 'PIL', 'cv2',
+    ],
+    win_no_prefer_redirects=False,
+    win_private_assemblies=False,
+    cipher=block_cipher,
+    noarchive=False,
+)
+
+pyz = PYZ(a.pure, a.zipped_data, cipher=block_cipher)
+
+exe = EXE(
+    pyz,
+    a.scripts,
+    [],
+    exclude_binaries=True,
+    name='local-transcription-backend',
+    debug=False,
+    bootloader_ignore_signals=False,
+    strip=False,
+    upx=True,
+    console=True,
+    disable_windowed_traceback=False,
+    argv_emulation=False,
+    target_arch=None,
+    codesign_identity=None,
+    entitlements_file=None,
+    icon='LocalTranscription.ico' if is_windows else None,
+)
+
+coll = COLLECT(
+    exe,
+    a.binaries,
+    a.zipfiles,
+    a.datas,
+    strip=False,
+    upx=True,
+    upx_exclude=[],
+    name='local-transcription-backend',
+)

local-transcription-headless.spec

@@ -38,6 +38,21 @@ datas = [
     (vad_assets_path, 'faster_whisper/assets'),
 ] + pvporcupine_data_files
 
+# Collect sounddevice's bundled PortAudio library (_sounddevice_data)
+try:
+    import sounddevice
+    sd_path = os.path.dirname(sounddevice.__file__)
+    sd_data = os.path.join(sd_path, '_sounddevice_data')
+    if os.path.exists(sd_data):
+        datas.append((sd_data, '_sounddevice_data'))
+        print(f"  + Collected sounddevice PortAudio data from {sd_data}")
+    sd_datas = collect_data_files('sounddevice')
+    if sd_datas:
+        datas += sd_datas
+        print(f"  + Collected {len(sd_datas)} sounddevice data files")
+except ImportError:
+    print("  - Warning: sounddevice not found")
+
 # Hidden imports -- NO PySide6/Qt needed for headless backend
 hiddenimports = [
     # Transcription engine
@@ -46,6 +61,7 @@ hiddenimports = [
     'faster_whisper.vad',
     'ctranslate2',
     'sounddevice',
+    '_sounddevice_data',
     'scipy',
     'scipy.signal',
     'numpy',

main_cli.py

@@ -90,7 +90,7 @@ class TranscriptionCLI:
             initial_prompt=self.config.get('transcription.initial_prompt', ''),
             no_log_file=True,
             input_device_index=audio_device,
-            user_name=user_name
+            app_config=self.config
         )
 
         # Set up callbacks

package.json

@@ -1,7 +1,7 @@
 {
   "name": "local-transcription",
   "private": true,
-  "version": "2.0.3",
+  "version": "2.0.20",
   "type": "module",
   "scripts": {
     "dev": "vite dev",

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "local-transcription"
-version = "1.0.4"
+version = "1.0.15"
 description = "A standalone desktop application for real-time speech-to-text transcription using Whisper models"
 readme = "README.md"
 requires-python = ">=3.9"

server/nodejs/server.js

@@ -703,6 +703,36 @@ app.post('/api/send', async (req, res) => {
     }
 });
 
+// Create room explicitly (no transcription needed)
+app.post('/api/create-room', async (req, res) => {
+    try {
+        const { room, passphrase } = req.body;
+
+        if (!room || !passphrase) {
+            return res.status(400).json({ error: 'Missing room or passphrase' });
+        }
+
+        // Check if room already exists
+        const existing = await loadRoom(room);
+        if (existing) {
+            const valid = await verifyPassphrase(room, passphrase);
+            if (!valid) {
+                return res.status(401).json({ error: 'Room exists with different passphrase' });
+            }
+            return res.json({ status: 'ok', room, created: false, message: 'Room already exists' });
+        }
+
+        // Create the room (verifyPassphrase creates it if it doesn't exist)
+        await verifyPassphrase(room, passphrase);
+
+        console.log(`[Room] Created room "${room}"`);
+        res.json({ status: 'ok', room, created: true });
+    } catch (err) {
+        console.error('Error in /api/create-room:', err);
+        res.status(500).json({ error: err.message });
+    }
+});
+
 // List transcriptions
 app.get('/api/list', async (req, res) => {
     try {

src-tauri/Cargo.lock

@@ -1881,7 +1881,7 @@ checksum = "92daf443525c4cce67b150400bc2316076100ce0b3686209eb8cf3c31612e6f0"
 
 [[package]]
 name = "local-transcription"
-version = "1.4.16"
+version = "2.0.12"
 dependencies = [
  "bytes",
  "chrono",

src-tauri/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "local-transcription"
-version = "2.0.3"
+version = "2.0.20"
 description = "Real-time speech-to-text transcription for streamers"
 authors = ["Local Transcription Contributors"]
 edition = "2021"

src-tauri/Entitlements.plist

@@ -0,0 +1,14 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+<plist version="1.0">
+<dict>
+    <key>com.apple.security.device.audio-input</key>
+    <true/>
+    <key>com.apple.security.network.client</key>
+    <true/>
+    <key>com.apple.security.network.server</key>
+    <true/>
+    <key>com.apple.security.cs.allow-unsigned-executable-memory</key>
+    <true/>
+</dict>
+</plist>

src-tauri/Info.plist

@@ -0,0 +1,8 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
+<plist version="1.0">
+<dict>
+    <key>NSMicrophoneUsageDescription</key>
+    <string>Local Transcription needs microphone access for real-time speech-to-text transcription.</string>
+</dict>
+</plist>

src-tauri/gen/schemas/capabilities.json

@@ -1 +1 @@
-{}
+{"default":{"identifier":"default","description":"Default permissions for the main window","local":true,"windows":["main"],"permissions":["core:default","core:event:default","core:event:allow-listen","core:event:allow-emit","shell:default","dialog:default","process:default"]}}

src-tauri/src/lib.rs

@@ -68,8 +68,31 @@ pub fn run() {
             sidecar::get_sidecar_port,
             sidecar::start_sidecar,
             sidecar::stop_sidecar,
+            sidecar::reset_sidecar,
             write_log,
         ])
-        .run(tauri::generate_context!())
+        .build(tauri::generate_context!())
-        .expect("error while running tauri application");
+        .expect("error while building tauri application")
+        .run(|app, event| {
+            match event {
+                tauri::RunEvent::Exit => {
+                    if let Some(state) = app.try_state::<sidecar::ManagedSidecar>() {
+                        if let Ok(mut mgr) = state.0.lock() {
+                            eprintln!("[app] Stopping sidecar on exit...");
+                            mgr.stop();
+                        }
+                    }
+                }
+                tauri::RunEvent::ExitRequested { .. } => {
+                    // Also stop sidecar on exit request (Cmd+Q on macOS)
+                    if let Some(state) = app.try_state::<sidecar::ManagedSidecar>() {
+                        if let Ok(mut mgr) = state.0.lock() {
+                            eprintln!("[app] Stopping sidecar on exit request...");
+                            mgr.stop();
+                        }
+                    }
+                }
+                _ => {}
+            }
+        });
 }

src-tauri/src/sidecar/mod.rs

@@ -554,18 +554,27 @@ impl SidecarManager {
     // -- private helpers -------------------------------------------------------
 
     fn build_dev_command(&self) -> Result<std::process::Command, String> {
-        let mut cmd = std::process::Command::new("python");
+        // Use `uv run` to ensure we use the project's venv, not system Python
-        cmd.args(["-u", "-m", "backend.main_headless"]); // -u = unbuffered
+        let mut cmd = std::process::Command::new("uv");
+        cmd.args(["run", "python", "-u", "-m", "backend.main_headless"]);
+
+        // Find the project root: try CARGO_MANIFEST_DIR first (set at compile time),
+        // then fall back to resource_dir parent chain
+        let manifest_dir = option_env!("CARGO_MANIFEST_DIR").map(std::path::PathBuf::from);
+        let project_root = manifest_dir
+            .as_ref()
+            .and_then(|d| d.parent()) // src-tauri -> project root
+            .or_else(|| {
+                DIRS.get()
+                    .and_then(|d| d.resource_dir.parent())
+                    .and_then(|p| p.parent())
+            });
 
-        // Try to find the project root (parent of src-tauri)
-        if let Some(dirs) = DIRS.get() {
-            let project_root = dirs
-                .resource_dir
-                .parent() // src-tauri
-                .and_then(|p| p.parent()); // project root
         if let Some(root) = project_root {
+            eprintln!("[sidecar] Dev mode: working dir = {}", root.display());
             cmd.current_dir(root);
-            }
+        } else {
+            eprintln!("[sidecar] Dev mode: WARNING - could not determine project root");
         }
 
         cmd.env("PYTHONUNBUFFERED", "1");
@@ -676,6 +685,42 @@ pub fn stop_sidecar(state: tauri::State<'_, ManagedSidecar>) -> Result<(), Strin
     Ok(())
 }
 
+/// Stop the running sidecar, delete its files and version marker.
+/// The next app launch will show the sidecar download prompt.
+#[tauri::command]
+pub fn reset_sidecar(state: tauri::State<'_, ManagedSidecar>) -> Result<(), String> {
+    // Stop the running sidecar first
+    {
+        let mut mgr = state
+            .0
+            .lock()
+            .map_err(|e| format!("Lock error: {e}"))?;
+        mgr.stop();
+    }
+
+    let data = data_dir();
+
+    // Delete the version file so check_sidecar returns false
+    let vf = version_file();
+    if vf.exists() {
+        std::fs::remove_file(&vf)
+            .map_err(|e| format!("Failed to delete version file: {e}"))?;
+    }
+
+    // Delete all sidecar directories
+    if let Ok(entries) = std::fs::read_dir(&data) {
+        for entry in entries.flatten() {
+            let name = entry.file_name().to_string_lossy().to_string();
+            if name.starts_with("sidecar-") && entry.path().is_dir() {
+                eprintln!("[sidecar] Removing {}", entry.path().display());
+                let _ = std::fs::remove_dir_all(entry.path());
+            }
+        }
+    }
+
+    Ok(())
+}
+
 // ---------------------------------------------------------------------------
 // Tests
 // ---------------------------------------------------------------------------

src-tauri/tauri.conf.json

@@ -1,6 +1,6 @@
 {
   "productName": "Local Transcription",
-  "version": "2.0.3",
+  "version": "2.0.20",
   "identifier": "net.anhonesthost.local-transcription",
   "build": {
     "frontendDist": "../dist",
@@ -33,7 +33,10 @@
       "icons/icon.icns",
       "icons/icon.ico",
       "icons/icon.png"
-    ]
+    ],
+    "windows": {
+      "digestAlgorithm": "sha256"
+    }
   },
   "plugins": {
     "shell": {

src/App.svelte

@@ -15,6 +15,7 @@
   let sidecarState = $state<SidecarState>("checking");
   let debugLog = $state("");
   let availableUpdate = $state("");
+  let appVersion = $state("");
 
   let obsDisplayUrl = $derived(backendStore.obsUrl);
   let syncDisplayUrl = $derived(backendStore.syncUrl);
@@ -108,6 +109,14 @@
   }
 
   onMount(() => {
+    // Get app version from Tauri
+    import("@tauri-apps/api/app").then(({ getVersion }) =>
+      getVersion().then((v) => { appVersion = v; })
+    ).catch(() => {
+      // Browser dev mode -- read from package.json or use fallback
+      appVersion = "dev";
+    });
+
     checkAndLaunchSidecar();
 
     return () => {
@@ -201,7 +210,7 @@
     <TranscriptionDisplay />
     <Controls />
 
-    <div class="version-label">v{backendStore.version}</div>
+    <div class="version-label">v{appVersion || backendStore.version}</div>
   </div>
 
   {#if showSettings}

src/lib/components/Controls.svelte

@@ -1,5 +1,6 @@
 <script lang="ts">
   import { backendStore } from "$lib/stores/backend";
+  import { configStore } from "$lib/stores/config";
   import { transcriptionStore } from "$lib/stores/transcriptions";
 
   let isTranscribing = $derived(backendStore.appState === "transcribing");
@@ -8,18 +9,39 @@
   );
   let isLoading = $state(false);
 
+  let remoteMode = $derived(configStore.config.remote.mode);
+  let byokApiKey = $derived(configStore.config.remote.byok_api_key);
+  let authToken = $derived(configStore.config.remote.auth_token);
+
+  let cloudConfigured = $derived(
+    remoteMode === "local" ||
+    (remoteMode === "byok" && byokApiKey.trim() !== "") ||
+    (remoteMode === "managed" && authToken.trim() !== "")
+  );
+
+  let errorMessage = $state("");
+
   async function toggleTranscription() {
     if (isLoading) return;
     isLoading = true;
+    errorMessage = "";
     try {
       if (isTranscribing) {
         await backendStore.apiPost("/api/stop");
       } else {
         await backendStore.apiPost("/api/start");
       }
-    } catch (err) {
+    } catch (err: unknown) {
-      console.error("Failed to toggle transcription:", err);
+      const msg = err instanceof Error ? err.message : String(err);
+      // Ignore "Already transcribing/not transcribing" -- just sync the state
+      if (!msg.includes("400")) {
+        console.error("Failed to toggle transcription:", msg);
+        errorMessage = msg;
+      }
     } finally {
+      // Always poll status to sync UI with actual backend state,
+      // even if the API call failed (e.g. "Already transcribing")
+      await backendStore.pollStatus();
       isLoading = false;
     }
   }
@@ -83,7 +105,7 @@
   <button
     class={isTranscribing ? "danger" : "primary"}
     onclick={toggleTranscription}
-    disabled={!isReady || isLoading}
+    disabled={!isReady || isLoading || !cloudConfigured}
   >
     {#if isLoading}
       ...
@@ -101,9 +123,43 @@
   <button onclick={saveTranscriptions} disabled={!backendStore.connected}>
     Save
   </button>
+
+  {#if errorMessage}
+    <span class="error-msg">{errorMessage}</span>
+  {/if}
+
+  {#if !cloudConfigured && isReady}
+    <div class="cloud-warning">
+      {#if remoteMode === "byok"}
+        <span>API key required. Get one at
+          <a href="https://console.deepgram.com" target="_blank" rel="noopener">console.deepgram.com</a>,
+          then enter it in Settings.</span>
+      {:else if remoteMode === "managed"}
+        <span>Login required. Open Settings to log in.</span>
+      {/if}
+    </div>
+  {/if}
 </div>
 
 <style>
+  .error-msg {
+    color: #f44336;
+    font-size: 12px;
+    margin-left: 8px;
+  }
+
+  .cloud-warning {
+    font-size: 12px;
+    color: #ff9800;
+    margin-left: 8px;
+    flex: 1;
+  }
+
+  .cloud-warning a {
+    color: #4fc3f7;
+    text-decoration: underline;
+  }
+
   .controls {
     display: flex;
     align-items: center;

src/lib/components/Settings.svelte

@@ -27,6 +27,10 @@
   let showTimestamps = $state(true);
   let fadeSeconds = $state(10);
   let maxLines = $state(100);
+  let fontSource = $state("System Font");
+  let fontFamily = $state("Courier");
+  let websafeFont = $state("Arial");
+  let googleFont = $state("Roboto");
   let fontSize = $state(12);
   let userColor = $state("#4CAF50");
   let textColor = $state("#FFFFFF");
@@ -40,8 +44,20 @@
   let byokApiKey = $state("");
   let managedEmail = $state("");
   let managedPassword = $state("");
+  let managedLoggedIn = $state(false);
   let autoCheckUpdates = $state(true);
 
+  let isCloudMode = $derived(remoteMode === "managed" || remoteMode === "byok");
+  let isCloudOnly = $derived(
+    computeDevices.some(d => d.id === "cloud")
+  );
+
+  // Room creation / join state
+  let shareCode = $state("");
+  let joinCode = $state("");
+  let roomCreating = $state(false);
+  let roomCreateMessage = $state("");
+
   let saving = $state(false);
   let saveMessage = $state("");
 
@@ -99,6 +115,10 @@
     showTimestamps = cfg.display.show_timestamps;
     fadeSeconds = cfg.display.fade_after_seconds;
     maxLines = cfg.display.max_lines;
+    fontSource = cfg.display.font_source ?? "System Font";
+    fontFamily = cfg.display.font_family ?? "Courier";
+    websafeFont = cfg.display.websafe_font ?? "Arial";
+    googleFont = cfg.display.google_font ?? "Roboto";
     fontSize = cfg.display.font_size;
     userColor = cfg.display.user_color;
     textColor = cfg.display.text_color;
@@ -112,6 +132,8 @@
     remoteMode = cfg.remote.mode;
     remoteServerUrl = cfg.remote.server_url;
     byokApiKey = cfg.remote.byok_api_key ?? "";
+    managedEmail = cfg.remote.email ?? "";
+    managedLoggedIn = !!(cfg.remote.auth_token && cfg.remote.email);
     autoCheckUpdates = cfg.updates.auto_check;
   });
 
@@ -174,6 +196,10 @@
         show_timestamps: showTimestamps,
         fade_after_seconds: fadeSeconds,
         max_lines: maxLines,
+        font_source: fontSource,
+        font_family: fontFamily,
+        websafe_font: websafeFont,
+        google_font: googleFont,
         font_size: fontSize,
         user_color: userColor,
         text_color: textColor,
@@ -187,7 +213,7 @@
       },
       remote: {
         mode: remoteMode,
-        server_url: remoteServerUrl,
+        server_url: remoteServerUrl || MANAGED_SERVER_URL,
         byok_api_key: byokApiKey,
       },
       updates: {
@@ -220,25 +246,155 @@
     }
   }
 
+  async function handleChangeSidecar() {
+    try {
+      const { invoke } = await import("@tauri-apps/api/core");
+      await invoke("reset_sidecar");
+      // Force a page reload which will re-trigger the setup flow
+      window.location.reload();
+    } catch (err) {
+      console.error("Failed to reset sidecar:", err);
+      saveMessage = `Error: ${err}`;
+    }
+  }
+
+  const MANAGED_SERVER_URL = "https://transcribe.shadowdao.com";
+
+  let loginMessage = $state("");
+
   async function handleManagedLogin() {
+    loginMessage = "";
     try {
       await backendStore.apiPost("/api/login", {
         email: managedEmail,
         password: managedPassword,
+        server_url: remoteServerUrl || MANAGED_SERVER_URL,
       });
+      loginMessage = "Logged in successfully!";
+      managedPassword = "";
+      managedLoggedIn = true;
+      await configStore.fetchConfig();
     } catch (err) {
       console.error("Login failed:", err);
+      loginMessage = "Login failed. Check your email and password.";
     }
   }
 
-  async function handleManagedRegister() {
+  async function handleManagedLogout() {
     try {
-      await backendStore.apiPost("/api/register", {
+      await configStore.updateConfig({
-        email: managedEmail,
+        remote: { auth_token: "", email: "" },
-        password: managedPassword,
       });
+      managedLoggedIn = false;
+      managedPassword = "";
+      loginMessage = "";
     } catch (err) {
-      console.error("Register failed:", err);
+      console.error("Logout failed:", err);
+      loginMessage = `Error: ${err}`;
+    }
+  }
+
+  const CAPTION_SERVER = "https://caption.shadowdao.com";
+
+  function generateRandomName(): string {
+    const adjectives = ['swift', 'bright', 'cosmic', 'electric', 'turbo', 'mega', 'ultra', 'super', 'hyper', 'alpha'];
+    const nouns = ['phoenix', 'dragon', 'tiger', 'falcon', 'comet', 'storm', 'blaze', 'thunder', 'frost', 'nebula'];
+    const num = Math.floor(Math.random() * 10000);
+    return `${adjectives[Math.floor(Math.random() * adjectives.length)]}-${nouns[Math.floor(Math.random() * nouns.length)]}-${num}`;
+  }
+
+  function generateRandomPassphrase(): string {
+    const chars = 'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789';
+    let result = '';
+    for (let i = 0; i < 16; i++) {
+      result += chars.charAt(Math.floor(Math.random() * chars.length));
+    }
+    return result;
+  }
+
+  function encodeShareCode(url: string, room: string, passphrase: string): string {
+    return btoa(JSON.stringify({ url, room, passphrase }));
+  }
+
+  function decodeShareCode(code: string): { url: string; room: string; passphrase: string } | null {
+    try {
+      const json = JSON.parse(atob(code.trim()));
+      if (json.url && json.room && json.passphrase) {
+        return json;
+      }
+      return null;
+    } catch {
+      return null;
+    }
+  }
+
+  async function handleCreateRoom() {
+    roomCreating = true;
+    roomCreateMessage = "";
+    shareCode = "";
+
+    const room = generateRandomName();
+    const passphrase = generateRandomPassphrase();
+    const serverSendUrl = `${CAPTION_SERVER}/api/send`;
+
+    try {
+      const resp = await fetch(`${CAPTION_SERVER}/api/create-room`, {
+        method: "POST",
+        headers: { "Content-Type": "application/json" },
+        body: JSON.stringify({ room, passphrase }),
+      });
+
+      if (!resp.ok) {
+        const err = await resp.json().catch(() => ({ error: "Request failed" }));
+        roomCreateMessage = `Error: ${err.error || resp.statusText}`;
+        return;
+      }
+
+      syncUrl = serverSendUrl;
+      syncRoom = room;
+      syncPassphrase = passphrase;
+      syncEnabled = true;
+
+      shareCode = encodeShareCode(serverSendUrl, room, passphrase);
+      roomCreateMessage = "Room created! Share the code below with others.";
+    } catch (err) {
+      roomCreateMessage = `Error: ${err instanceof Error ? err.message : String(err)}`;
+    } finally {
+      roomCreating = false;
+    }
+  }
+
+  function handleJoinRoom() {
+    const decoded = decodeShareCode(joinCode);
+    if (!decoded) {
+      roomCreateMessage = "Invalid share code. Please check and try again.";
+      return;
+    }
+    syncUrl = decoded.url;
+    syncRoom = decoded.room;
+    syncPassphrase = decoded.passphrase;
+    syncEnabled = true;
+    joinCode = "";
+    roomCreateMessage = "Room joined! Fields have been auto-filled.";
+  }
+
+  async function handleShareCurrentRoom() {
+    const code = encodeShareCode(syncUrl, syncRoom, syncPassphrase);
+    shareCode = code;
+    try {
+      await navigator.clipboard.writeText(code);
+      roomCreateMessage = "Share code copied to clipboard!";
+    } catch {
+      roomCreateMessage = "Share code generated. Copy it from the field below.";
+    }
+  }
+
+  async function copyShareCode() {
+    try {
+      await navigator.clipboard.writeText(shareCode);
+      roomCreateMessage = "Share code copied to clipboard!";
+    } catch {
+      roomCreateMessage = "Failed to copy. Please select and copy manually.";
     }
   }
 
@@ -303,7 +459,100 @@
         </div>
       </section>
 
-      <!-- Transcription Settings -->
+      <!-- Remote Transcription (moved up for cloud-first UX) -->
+      <section class="settings-section">
+        <h3>Transcription Mode</h3>
+        <div class="radio-group">
+          <label>
+            <input
+              type="radio"
+              name="remote-mode"
+              value="byok"
+              bind:group={remoteMode}
+            />
+            Cloud (Deepgram)
+          </label>
+          <label>
+            <input
+              type="radio"
+              name="remote-mode"
+              value="managed"
+              bind:group={remoteMode}
+            />
+            Managed Service
+          </label>
+          {#if !isCloudOnly}
+          <label>
+            <input
+              type="radio"
+              name="remote-mode"
+              value="local"
+              bind:group={remoteMode}
+            />
+            Local (Whisper)
+          </label>
+          {/if}
+        </div>
+        {#if remoteMode === "byok"}
+          <div class="field">
+            <label for="byok-key">Deepgram API Key</label>
+            <input
+              id="byok-key"
+              type="password"
+              bind:value={byokApiKey}
+              placeholder="Enter your Deepgram API key"
+            />
+            <p style="font-size: 11px; color: var(--text-muted); margin-top: 4px;">
+              Get a key at <a href="https://console.deepgram.com" target="_blank" rel="noopener" style="color: var(--accent-blue);">console.deepgram.com</a>
+            </p>
+          </div>
+        {/if}
+        {#if remoteMode === "managed"}
+          <div class="managed-auth">
+            {#if managedLoggedIn}
+              <p style="font-size: 13px; margin: 0 0 8px;">
+                <span style="color: var(--accent-green, #4CAF50);">✓ Logged in</span>
+                as <strong>{managedEmail}</strong>
+              </p>
+              <div class="auth-buttons">
+                <button onclick={handleManagedLogout}>Log out</button>
+              </div>
+            {:else}
+              <div class="field">
+                <label for="managed-email">Email</label>
+                <input
+                  id="managed-email"
+                  type="email"
+                  bind:value={managedEmail}
+                  placeholder="email@example.com"
+                />
+              </div>
+              <div class="field">
+                <label for="managed-password">Password</label>
+                <input
+                  id="managed-password"
+                  type="password"
+                  bind:value={managedPassword}
+                />
+              </div>
+              <div class="auth-buttons">
+                <button onclick={handleManagedLogin}>Login</button>
+              </div>
+              <p style="font-size: 11px; color: var(--text-muted); margin-top: 8px;">
+                Don't have an account? <a href="https://transcribe.shadowdao.com/register.html" target="_blank" rel="noopener" style="color: var(--accent-blue);">Sign up here</a>
+              </p>
+            {/if}
+            {#if loginMessage}
+              <p style="font-size: 12px; margin-top: 6px; color: {loginMessage.startsWith('Logged') ? 'var(--accent-green, #4CAF50)' : 'var(--accent-red, #f44336)'};">
+                {loginMessage}
+              </p>
+            {/if}
+          </div>
+        {/if}
+      </section>
+
+      {#if !isCloudMode}
+      <!-- Transcription Settings (local Whisper only) -->
       <section class="settings-section">
         <h3>Transcription Settings</h3>
         <div class="field">
@@ -449,6 +698,7 @@
           />
         </div>
       </section>
+      {/if}
 
       <!-- Display Settings -->
       <section class="settings-section">
@@ -485,6 +735,95 @@
             bind:value={maxLines}
           />
         </div>
+        <div class="field">
+          <label for="font-source">Font Source</label>
+          <select id="font-source" bind:value={fontSource}>
+            <option value="System Font">System Font</option>
+            <option value="Web-Safe">Web-Safe</option>
+            <option value="Google Font">Google Font</option>
+          </select>
+        </div>
+        {#if fontSource === "System Font"}
+          <div class="field">
+            <label for="font-family">System Font Family</label>
+            <input id="font-family" type="text" bind:value={fontFamily} placeholder="Courier" />
+          </div>
+        {/if}
+        {#if fontSource === "Web-Safe"}
+          <div class="field">
+            <label for="websafe-font">Web-Safe Font</label>
+            <select id="websafe-font" bind:value={websafeFont}>
+              <option value="Arial">Arial</option>
+              <option value="Arial Black">Arial Black</option>
+              <option value="Comic Sans MS">Comic Sans MS</option>
+              <option value="Courier New">Courier New</option>
+              <option value="Georgia">Georgia</option>
+              <option value="Impact">Impact</option>
+              <option value="Lucida Console">Lucida Console</option>
+              <option value="Lucida Sans Unicode">Lucida Sans Unicode</option>
+              <option value="Palatino Linotype">Palatino Linotype</option>
+              <option value="Tahoma">Tahoma</option>
+              <option value="Times New Roman">Times New Roman</option>
+              <option value="Trebuchet MS">Trebuchet MS</option>
+              <option value="Verdana">Verdana</option>
+            </select>
+          </div>
+        {/if}
+        {#if fontSource === "Google Font"}
+          <div class="field">
+            <label for="google-font">Google Font</label>
+            <select id="google-font" bind:value={googleFont}>
+              <optgroup label="Sans Serif">
+                <option value="Roboto">Roboto</option>
+                <option value="Open Sans">Open Sans</option>
+                <option value="Lato">Lato</option>
+                <option value="Montserrat">Montserrat</option>
+                <option value="Poppins">Poppins</option>
+                <option value="Nunito">Nunito</option>
+                <option value="Raleway">Raleway</option>
+                <option value="Ubuntu">Ubuntu</option>
+                <option value="Rubik">Rubik</option>
+                <option value="Work Sans">Work Sans</option>
+                <option value="Inter">Inter</option>
+                <option value="Outfit">Outfit</option>
+                <option value="Quicksand">Quicksand</option>
+                <option value="Comfortaa">Comfortaa</option>
+                <option value="Varela Round">Varela Round</option>
+              </optgroup>
+              <optgroup label="Serif">
+                <option value="Playfair Display">Playfair Display</option>
+                <option value="Merriweather">Merriweather</option>
+                <option value="Lora">Lora</option>
+                <option value="PT Serif">PT Serif</option>
+                <option value="Crimson Text">Crimson Text</option>
+              </optgroup>
+              <optgroup label="Monospace">
+                <option value="Roboto Mono">Roboto Mono</option>
+                <option value="Source Code Pro">Source Code Pro</option>
+                <option value="Fira Code">Fira Code</option>
+                <option value="JetBrains Mono">JetBrains Mono</option>
+                <option value="IBM Plex Mono">IBM Plex Mono</option>
+              </optgroup>
+              <optgroup label="Display">
+                <option value="Bebas Neue">Bebas Neue</option>
+                <option value="Oswald">Oswald</option>
+                <option value="Righteous">Righteous</option>
+                <option value="Bangers">Bangers</option>
+                <option value="Permanent Marker">Permanent Marker</option>
+              </optgroup>
+              <optgroup label="Handwriting">
+                <option value="Pacifico">Pacifico</option>
+                <option value="Lobster">Lobster</option>
+                <option value="Dancing Script">Dancing Script</option>
+                <option value="Caveat">Caveat</option>
+                <option value="Satisfy">Satisfy</option>
+              </optgroup>
+            </select>
+            <p style="font-size: 11px; color: var(--text-muted); margin-top: 4px;">
+              Browse more at <a href="https://fonts.google.com" target="_blank" rel="noopener" style="color: var(--accent-blue);">fonts.google.com</a>
+            </p>
+          </div>
+        {/if}
         <div class="field">
           <label for="font-size">Font Size: {fontSize}px</label>
           <input
@@ -515,11 +854,11 @@
         </div>
       </section>
 
-      <!-- Server Sync -->
+      <!-- Server Sync (Shared Captions) -->
       <section class="settings-section">
-        <h3>Server Sync</h3>
+        <h3>Shared Captions</h3>
         <div class="field-row">
-          <label for="sync-enabled">Enable Server Sync</label>
+          <label for="sync-enabled">Enable Shared Captions</label>
           <input
             id="sync-enabled"
             type="checkbox"
@@ -527,13 +866,57 @@
           />
         </div>
         {#if syncEnabled}
+          <div class="room-actions">
+            <div class="room-buttons-row">
+              <button
+                onclick={handleCreateRoom}
+                disabled={roomCreating}
+                class="secondary"
+              >
+                {roomCreating ? "Creating..." : "Create Room"}
+              </button>
+              <button
+                onclick={handleShareCurrentRoom}
+                disabled={!syncUrl.trim() || !syncRoom.trim() || !syncPassphrase.trim()}
+                class="secondary"
+              >
+                Share Current Room
+              </button>
+            </div>
+            <div class="join-row">
+              <input
+                type="text"
+                bind:value={joinCode}
+                placeholder="Paste share code to join"
+                class="join-input"
+              />
+              <button onclick={handleJoinRoom} disabled={!joinCode.trim()} class="secondary">
+                Join
+              </button>
+            </div>
+          </div>
+
+          {#if roomCreateMessage}
+            <p class="room-message" class:error={roomCreateMessage.startsWith("Error")}>{roomCreateMessage}</p>
+          {/if}
+
+          {#if shareCode}
+            <div class="share-code-box">
+              <label>Share Code</label>
+              <div class="share-code-row">
+                <input type="text" value={shareCode} readonly class="share-code-input" />
+                <button onclick={copyShareCode} class="secondary">Copy</button>
+              </div>
+            </div>
+          {/if}
+
           <div class="field">
             <label for="sync-url">Server URL</label>
             <input
               id="sync-url"
               type="url"
               bind:value={syncUrl}
-              placeholder="http://localhost:3000/api/send"
+              placeholder="https://caption.shadowdao.com/api/send"
             />
           </div>
           <div class="field">
@@ -551,90 +934,6 @@
         {/if}
       </section>
 
-      <!-- Remote Transcription -->
-      <section class="settings-section">
-        <h3>Remote Transcription</h3>
-        <div class="radio-group">
-          <label>
-            <input
-              type="radio"
-              name="remote-mode"
-              value="local"
-              bind:group={remoteMode}
-            />
-            Local
-          </label>
-          <label>
-            <input
-              type="radio"
-              name="remote-mode"
-              value="managed"
-              bind:group={remoteMode}
-            />
-            Managed
-          </label>
-          <label>
-            <input
-              type="radio"
-              name="remote-mode"
-              value="byok"
-              bind:group={remoteMode}
-            />
-            BYOK (Bring Your Own Key)
-          </label>
-        </div>
-        {#if remoteMode === "managed"}
-          <div class="field">
-            <label for="remote-url">Server URL</label>
-            <input
-              id="remote-url"
-              type="url"
-              bind:value={remoteServerUrl}
-              placeholder="wss://your-proxy.com"
-            />
-          </div>
-        {/if}
-        {#if remoteMode === "byok"}
-          <div class="field">
-            <label for="byok-key">Deepgram API Key</label>
-            <input
-              id="byok-key"
-              type="password"
-              bind:value={byokApiKey}
-              placeholder="Enter your Deepgram API key"
-            />
-            <p style="font-size: 11px; color: var(--text-muted); margin-top: 4px;">
-              Get a key at <a href="https://console.deepgram.com" target="_blank" rel="noopener" style="color: var(--accent-blue);">console.deepgram.com</a>
-            </p>
-          </div>
-        {/if}
-        {#if remoteMode === "managed"}
-          <div class="managed-auth">
-            <div class="field">
-              <label for="managed-email">Email</label>
-              <input
-                id="managed-email"
-                type="email"
-                bind:value={managedEmail}
-                placeholder="email@example.com"
-              />
-            </div>
-            <div class="field">
-              <label for="managed-password">Password</label>
-              <input
-                id="managed-password"
-                type="password"
-                bind:value={managedPassword}
-              />
-            </div>
-            <div class="auth-buttons">
-              <button onclick={handleManagedLogin}>Login</button>
-              <button onclick={handleManagedRegister}>Register</button>
-            </div>
-          </div>
-        {/if}
-      </section>
-
       <!-- Updates -->
       <section class="settings-section">
         <h3>Updates</h3>
@@ -648,6 +947,17 @@
         </div>
         <button onclick={handleCheckUpdates}>Check Now</button>
       </section>
+
+      <!-- Transcription Engine -->
+      <section class="settings-section">
+        <h3>Transcription Engine</h3>
+        <p style="font-size: 12px; color: var(--text-secondary); margin-bottom: 12px;">
+          Switch between local (Whisper) and cloud (Deepgram) transcription engines.
+          This will stop the current engine, remove the downloaded files, and restart
+          with the new engine selection.
+        </p>
+        <button class="danger-btn" onclick={handleChangeSidecar}>Change Transcription Engine</button>
+      </section>
     </div>
 
     <div class="settings-footer">
@@ -818,4 +1128,90 @@
   .save-message.error {
     color: #f44336;
   }
+
+  .room-actions {
+    display: flex;
+    flex-direction: column;
+    gap: 8px;
+    margin-bottom: 12px;
+  }
+
+  .room-buttons-row {
+    display: flex;
+    gap: 8px;
+  }
+
+  .join-row {
+    display: flex;
+    gap: 8px;
+  }
+
+  .join-input {
+    flex: 1;
+  }
+
+  .room-message {
+    font-size: 12px;
+    color: #4CAF50;
+    margin-bottom: 8px;
+  }
+
+  .room-message.error {
+    color: #f44336;
+  }
+
+  .share-code-box {
+    margin-bottom: 12px;
+  }
+
+  .share-code-box label {
+    display: block;
+    margin-bottom: 4px;
+    font-size: 12px;
+    color: var(--text-secondary);
+  }
+
+  .share-code-row {
+    display: flex;
+    gap: 8px;
+  }
+
+  .share-code-input {
+    flex: 1;
+    font-size: 11px;
+    font-family: monospace;
+  }
+
+  .secondary {
+    background: transparent;
+    border: 1px solid var(--border-color);
+    color: var(--text-primary);
+    padding: 6px 12px;
+    border-radius: 6px;
+    cursor: pointer;
+    font-size: 13px;
+  }
+
+  .secondary:hover {
+    background: var(--bg-tertiary);
+  }
+
+  .secondary:disabled {
+    opacity: 0.5;
+    cursor: not-allowed;
+  }
+
+  .danger-btn {
+    background: transparent;
+    border: 1px solid var(--accent-red, #f44336);
+    color: var(--accent-red, #f44336);
+    padding: 8px 16px;
+    border-radius: 6px;
+    cursor: pointer;
+    font-size: 13px;
+  }
+
+  .danger-btn:hover {
+    background: rgba(244, 67, 54, 0.1);
+  }
 </style>

src/lib/components/SidecarSetup.svelte

@@ -36,11 +36,12 @@
 
     try {
       // Listen for progress events from the Tauri backend
-      unlisten = await listen<{ progress: number; message: string }>(
+      unlisten = await listen<{ downloaded: number; total: number; phase: string; message: string }>(
         "sidecar-download-progress",
         (event) => {
-          progress = event.payload.progress;
+          const { downloaded, total, message } = event.payload;
-          progressMessage = event.payload.message;
+          progress = total > 0 ? (downloaded / total) * 100 : 0;
+          progressMessage = message;
         }
       );
 
@@ -84,11 +85,29 @@
 
     {#if setupState === "choose"}
       <p class="setup-description">
-        The app needs to download its transcription engine before you can start.
+        Choose a transcription engine. You can change this later in Settings.
-        Choose the version that best fits your hardware.
       </p>
 
       <div class="variant-options">
+        <label class="variant-option" class:selected={variant === "cloud"}>
+          <input
+            type="radio"
+            name="variant"
+            value="cloud"
+            bind:group={variant}
+          />
+          <div class="variant-info">
+            <span class="variant-name">Cloud (Deepgram)</span>
+            <span class="variant-desc">~50 MB download</span>
+            <span class="variant-detail">
+              Fast, accurate streaming transcription via Deepgram's servers.
+              Requires internet and a Deepgram API key.
+              Best for most users — low resource usage, works on any hardware.
+            </span>
+            <span class="variant-tag recommended">Recommended</span>
+          </div>
+        </label>
+
         <label class="variant-option" class:selected={variant === "cpu"}>
           <input
             type="radio"
@@ -97,23 +116,16 @@
             bind:group={variant}
           />
           <div class="variant-info">
-            <span class="variant-name">Standard (CPU)</span>
+            <span class="variant-name">Local - CPU</span>
-            <span class="variant-desc">Works on all computers (~500 MB download)</span>
+            <span class="variant-desc">~500 MB download</span>
+            <span class="variant-detail">
+              Runs Whisper AI models locally on your CPU. No internet needed
+              after download. Good for privacy or offline use, but slower and
+              uses more system resources than cloud.
+            </span>
           </div>
         </label>
 
-        <label class="variant-option" class:selected={variant === "cuda"}>
-          <input
-            type="radio"
-            name="variant"
-            value="cuda"
-            bind:group={variant}
-          />
-          <div class="variant-info">
-            <span class="variant-name">GPU Accelerated (CUDA)</span>
-            <span class="variant-desc">Faster transcription with NVIDIA GPU (~2 GB download)</span>
-          </div>
-        </label>
       </div>
 
       <button class="download-btn" onclick={startDownload}>
@@ -260,6 +272,30 @@
     color: #888;
   }
 
+  .variant-detail {
+    font-size: 11px;
+    color: #666;
+    line-height: 1.4;
+    margin-top: 2px;
+  }
+
+  .variant-tag {
+    display: inline-block;
+    font-size: 10px;
+    font-weight: 600;
+    text-transform: uppercase;
+    letter-spacing: 0.5px;
+    padding: 2px 6px;
+    border-radius: 3px;
+    margin-top: 4px;
+    width: fit-content;
+  }
+
+  .variant-tag.recommended {
+    background: rgba(76, 175, 80, 0.15);
+    color: #4CAF50;
+  }
+
   .download-btn {
     display: block;
     width: 100%;

src/lib/stores/backend.svelte.ts

@@ -19,6 +19,7 @@ interface BackendState {
   wsConnection: WebSocket | null;
   version: string;
   lastError: string;
+  isCloudOnly: boolean;
 }
 
 let state = $state<BackendState>({
@@ -30,6 +31,7 @@ let state = $state<BackendState>({
   wsConnection: null,
   version: "1.4.0",
   lastError: "",
+  isCloudOnly: false,
 });
 
 let reconnectTimer: ReturnType<typeof setTimeout> | null = null;
@@ -72,6 +74,9 @@ async function pollStatus() {
       if (data.version) {
         state.version = data.version;
       }
+      if (data.is_cloud_only !== undefined) {
+        state.isCloudOnly = data.is_cloud_only;
+      }
     }
   } catch {
     // API not ready yet, will retry
@@ -285,6 +290,9 @@ export const backendStore = {
   get lastError() {
     return state.lastError;
   },
+  get isCloudOnly() {
+    return state.isCloudOnly;
+  },
   get apiBaseUrl() {
     return `http://localhost:${state.port}`;
   },
@@ -302,6 +310,7 @@ export const backendStore = {
   setPort,
   connect: connectWebSocket,
   disconnect,
+  pollStatus,
   apiUrl,
   apiFetch,
   apiGet,

src/lib/stores/config.svelte.ts

@@ -65,6 +65,7 @@ export interface AppConfig {
     mode: string;
     server_url: string;
     auth_token: string;
+    email: string;
     byok_api_key: string;
     deepgram_model: string;
     language: string;
@@ -107,7 +108,7 @@ function getDefaultConfig(): AppConfig {
     },
     server_sync: {
       enabled: false,
-      url: "http://localhost:3000/api/send",
+      url: "",
       room: "default",
       passphrase: "",
     },
@@ -128,9 +129,10 @@ function getDefaultConfig(): AppConfig {
     },
     web_server: { port: 8080, host: "127.0.0.1" },
     remote: {
-      mode: "local",
+      mode: "byok",
       server_url: "",
       auth_token: "",
+      email: "",
       byok_api_key: "",
       deepgram_model: "nova-2",
       language: "en-US",

version.py

@@ -1,7 +1,7 @@
 """Version information for Local Transcription."""
 
-__version__ = "2.0.3"
+__version__ = "2.0.20"
-__version_info__ = (2, 0, 3)
+__version_info__ = (2, 0, 20)
 
 # Version history:
 # 1.4.0 - Auto-update feature: