diff --git a/.agents/scripts/speech-to-speech-helper.sh b/.agents/scripts/speech-to-speech-helper.sh
index 96f8406ff..4f12d39b4 100755
--- a/.agents/scripts/speech-to-speech-helper.sh
+++ b/.agents/scripts/speech-to-speech-helper.sh
@@ -212,7 +212,11 @@ cmd_start() {
         case "$gpu" in
             mps)  mode="local-mac" ;;
             cuda) mode="cuda" ;;
-            *)    mode="cuda" ;;
+            cpu)
+                mode="server"
+                print_warning "CPU-only host detected; defaulting to --server mode"
+                ;;
+            *)    mode="server" ;;
         esac
         print_info "Auto-detected mode: $mode"
     fi
@@ -314,7 +318,7 @@ cmd_docker_start() {
     fi
 
     print_info "Starting with Docker..."
-    (cd "$S2S_DIR" && docker compose up -d) || exit
+    (cd "$S2S_DIR" && docker compose up -d) || return 1
     print_success "Docker containers started"
     print_info "Ports: ${DEFAULT_RECV_PORT} (recv), ${DEFAULT_SEND_PORT} (send)"
     return 0
@@ -388,10 +392,10 @@ cmd_stop() {
     fi
 
     # Stop Docker if running
-    if [[ -f "${S2S_DIR}/docker-compose.yml" ]]; then
+    if command -v docker &>/dev/null && [[ -f "${S2S_DIR}/docker-compose.yml" ]]; then
         if docker compose -f "${S2S_DIR}/docker-compose.yml" ps --quiet 2>/dev/null | grep -q .; then
             print_info "Stopping Docker containers..."
-            (cd "$S2S_DIR" && docker compose down) || exit
+            (cd "$S2S_DIR" && docker compose down) || return 1
             print_success "Docker containers stopped"
         fi
     fi
@@ -543,7 +547,7 @@ cmd_help() {
     echo "  $0 start --local-mac"
     echo "  $0 start --cuda --language auto --background"
     echo "  $0 start --server"
-    echo "  $0 client --host 192.168.1.100"
+    echo "  $0 client --host <server-ip>"
     echo "  $0 start \$($0 config low-latency)"
     echo "  $0 stop"
     echo ""
diff --git a/.agents/tools/voice/speech-to-speech.md b/.agents/tools/voice/speech-to-speech.md
index d14666fd0..73865bc90 100644
--- a/.agents/tools/voice/speech-to-speech.md
+++ b/.agents/tools/voice/speech-to-speech.md
@@ -81,6 +81,8 @@ Model selection: `--stt_model_name <model>` (any Whisper checkpoint on HF Hub)
 
 Model selection: `--lm_model_name <model>` or `--mlx_lm_model_name <model>`
 
+API keys: Store `OPENAI_API_KEY` via `aidevops secret set OPENAI_API_KEY` (gopass encrypted) or in `~/.config/aidevops/credentials.sh` (600 permissions). See `tools/credentials/api-key-setup.md`.
+
 ### TTS (Text to Speech)
 
 | Implementation | Flag | Best For |
@@ -177,7 +179,7 @@ python -m unidic download
 ### Requirements
 
 - Python 3.10+
-- PyTorch 2.4+ (CUDA) or 2.10+ (macOS)
+- PyTorch 2.4+ (CUDA and macOS)
 - `uv` package manager (recommended)
 - CUDA 12.1+ (for GPU) or Apple Silicon (for MPS)
 - `sounddevice` for local audio I/O