diff --git a/stacks/fish-cpp/Dockerfile b/stacks/fish-cpp/Dockerfile
index 2dfd35b..b1afff3 100644
--- a/stacks/fish-cpp/Dockerfile
+++ b/stacks/fish-cpp/Dockerfile
@@ -17,7 +17,12 @@ ENV DEBIAN_FRONTEND=noninteractive
 
 RUN apt-get update && apt-get install -y --no-install-recommends \
       git ca-certificates cmake ninja-build build-essential pkg-config \
+      libvulkan-dev glslc \
     && rm -rf /var/lib/apt/lists/*
+# libvulkan-dev = Vulkan headers + loader for build-time linking.
+# glslc = GLSL→SPIR-V compiler; GGML's Vulkan backend compiles its
+# shaders with this at build time. Without it, the cmake configure
+# step disables Vulkan silently.
 
 WORKDIR /src
 RUN git clone --recurse-submodules https://github.com/rodrigomatta/s2.cpp.git \
@@ -38,7 +43,14 @@ WORKDIR /src/s2.cpp
 RUN ln -s /usr/local/cuda/lib64/stubs/libcuda.so /usr/local/lib/libcuda.so \
     && ln -s /usr/local/cuda/lib64/stubs/libcuda.so /usr/local/lib/libcuda.so.1 \
     && ldconfig
-RUN cmake -G Ninja -B build -DCMAKE_BUILD_TYPE=Release -DS2_CUDA=ON \
+# Build with BOTH CUDA and Vulkan backends — runtime selects via the
+# entrypoint (-c <id> for CUDA, -v <id> for Vulkan). We've established
+# that s2.cpp's CUDA path silently falls back to CPU on this model
+# (alpha software, incomplete op coverage); Vulkan is the README's
+# canonical example so it likely has more complete op coverage on
+# fish-speech. Keep both built so we can flip without rebuilding.
+RUN cmake -G Ninja -B build -DCMAKE_BUILD_TYPE=Release \
+      -DS2_CUDA=ON -DS2_VULKAN=ON \
     && cmake --build build --parallel $(nproc) --target s2
 
 # ── Stage 2: runtime — slim image with the binary + python shim ────────
@@ -51,9 +63,13 @@ ENV DEBIAN_FRONTEND=noninteractive \
 RUN apt-get update && apt-get install -y --no-install-recommends \
       python3 python3-pip python3-venv tini \
       libgomp1 curl \
+      libvulkan1 \
     && rm -rf /var/lib/apt/lists/*
-# libgomp1 = GNU OpenMP runtime — required by the s2 binary.
-# curl    = entrypoint uses it to wait for s2 server to bind :3030.
+# libgomp1   = GNU OpenMP runtime — required by the s2 binary.
+# curl       = entrypoint uses it to wait for s2 server to bind :3030.
+# libvulkan1 = Vulkan loader. The NVIDIA Vulkan ICD itself comes from
+#              the host driver via NVIDIA container runtime when
+#              NVIDIA_DRIVER_CAPABILITIES includes "graphics".
 
 # Pull the shim deps into an isolated venv so we don't fight system pip.
 RUN python3 -m venv /opt/venv
@@ -64,6 +80,7 @@ RUN pip install --no-cache-dir 'fastapi>=0.115' 'uvicorn[standard]>=0.30' 'pydan
 COPY --from=builder /src/s2.cpp/build/s2 /usr/local/bin/s2
 COPY --from=builder /src/s2.cpp/build/ggml/src/libggml*.so /usr/local/lib/
 COPY --from=builder /src/s2.cpp/build/ggml/src/ggml-cuda/libggml-cuda.so /usr/local/lib/
+COPY --from=builder /src/s2.cpp/build/ggml/src/ggml-vulkan/libggml-vulkan.so /usr/local/lib/
 RUN ldconfig
 
 WORKDIR /app
diff --git a/stacks/fish-cpp/compose.yaml b/stacks/fish-cpp/compose.yaml
index 293a706..94ea93b 100644
--- a/stacks/fish-cpp/compose.yaml
+++ b/stacks/fish-cpp/compose.yaml
@@ -35,9 +35,18 @@ services:
       - "${FISH_CPP_BIND:-0.0.0.0}:${FISH_CPP_PORT}:8000"
     environment:
       - NVIDIA_VISIBLE_DEVICES=${FISH_CPP_GPU_DEVICES:-1}
+      # Default nvidia-container-toolkit only mounts compute libs.
+      # Vulkan needs the graphics ICD too (libGLX_nvidia, vulkan ICD
+      # JSON). Without this, Vulkan init in the container fails with
+      # "no Vulkan ICD" even though the GPU is present.
+      - NVIDIA_DRIVER_CAPABILITIES=compute,utility,graphics
       - FISH_CPP_MODEL=${FISH_CPP_MODEL:-s2-pro-q6_k.gguf}
       - FISH_CPP_TOKENIZER=tokenizer.json
       - FISH_CPP_DEVICE=0
+      # Backend selection for the s2 server in entrypoint.sh:
+      #   cuda    — -c <id>; alpha CUDA path, GPU 0% util on fish-speech (broken)
+      #   vulkan  — -v <id>; README's canonical example, more battle-tested
+      - FISH_CPP_BACKEND=${FISH_CPP_BACKEND:-vulkan}
     volumes:
       - ${FISH_CPP_WEIGHTS_DIR}:/weights:ro
       - ${FISH_CPP_REFERENCE_DIR}:/references:ro
diff --git a/stacks/fish-cpp/entrypoint.sh b/stacks/fish-cpp/entrypoint.sh
index f9060d8..2a1cb3d 100644
--- a/stacks/fish-cpp/entrypoint.sh
+++ b/stacks/fish-cpp/entrypoint.sh
@@ -11,18 +11,26 @@ set -euo pipefail
 MODEL_PATH="${WEIGHTS_DIR:-/weights}/${FISH_CPP_MODEL:-s2-pro-q6_k.gguf}"
 TOKENIZER_PATH="${WEIGHTS_DIR:-/weights}/${FISH_CPP_TOKENIZER:-tokenizer.json}"
 DEVICE="${FISH_CPP_DEVICE:-0}"
+BACKEND="${FISH_CPP_BACKEND:-vulkan}"
+
+case "$BACKEND" in
+  cuda)   BACKEND_FLAG=(-c "$DEVICE") ;;
+  vulkan) BACKEND_FLAG=(-v "$DEVICE") ;;
+  cpu)    BACKEND_FLAG=() ;;  # no flag → s2 stays on CPU
+  *)      echo "[entrypoint] unknown FISH_CPP_BACKEND=$BACKEND (use cuda/vulkan/cpu)" >&2; exit 2 ;;
+esac
 
 if [ ! -f "$MODEL_PATH" ];     then echo "missing model: $MODEL_PATH" >&2;     exit 1; fi
 if [ ! -f "$TOKENIZER_PATH" ]; then echo "missing tokenizer: $TOKENIZER_PATH" >&2; exit 1; fi
 
-# Background s2 server. Loads model on GPU once, then accepts
-# multipart POSTs on localhost:3030/generate.
-echo "[entrypoint] starting s2 server on :3030 with CUDA device $DEVICE"
+# Background s2 server. Loads model on GPU once (per backend), then
+# accepts multipart POSTs on localhost:3030/generate.
+echo "[entrypoint] starting s2 server on :3030 with $BACKEND backend, device $DEVICE"
 /usr/local/bin/s2 \
   --server -H 127.0.0.1 -P 3030 \
   -m "$MODEL_PATH" \
   -t "$TOKENIZER_PATH" \
-  -c "$DEVICE" &
+  "${BACKEND_FLAG[@]}" &
 S2_PID=$!
 
 # Wait for s2 to bind 3030 before starting the shim. Avoids the