diff --git a/Dockerfile b/Dockerfile
index 0deddce349..d6c3bfad6f 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -27,10 +27,8 @@ RUN apk add --no-cache \
     npm \
     libsndfile
 
-ENV PRISMA_BINARY_CACHE_DIR=/app/.cache/prisma-python/binaries \
-    UV_PROJECT_ENVIRONMENT=/app/.venv \
+ENV UV_PROJECT_ENVIRONMENT=/app/.venv \
     UV_LINK_MODE=copy \
-    XDG_CACHE_HOME=/app/.cache \
     PATH="/app/.venv/bin:${PATH}"
 
 # Copy dependency metadata first for layer caching
@@ -97,18 +95,15 @@ WORKDIR /app
 ENV PATH="/app/.venv/bin:${PATH}"
 
 COPY --from=builder /app /app
+# Prisma binaries live in $HOME/.cache (default prisma-python location),
+# which is /root/.cache here. Copy them from the builder so they survive
+# deployments that volume-mount /app/.cache (e.g. readOnlyRootFilesystem
+# + emptyDir) — otherwise the mount would shadow the baked-in query engine.
+COPY --from=builder /root/.cache /root/.cache
 
 RUN find /app/.venv -type f -path "*/tornado/test/*" -delete && \
     find /app/.venv -type d -path "*/tornado/test" -delete
 
-# Regenerate the Prisma client in the runtime stage so the baked-in
-# BINARY_PATHS resolve to a location outside /app. Users with volume mounts
-# that shadow /app/.cache (e.g. readOnlyRootFilesystem + emptyDir) would
-# otherwise lose access to the pre-downloaded query engine at runtime.
-# Drop the builder's /app/.cache afterwards — it's stale and adds ~800 MB
-# the runtime doesn't use.
-RUN rm -rf /app/.cache && prisma generate --schema=./schema.prisma
-
 EXPOSE 4000/tcp
 
 COPY docker/supervisord.conf /etc/supervisord.conf
diff --git a/docker/Dockerfile.database b/docker/Dockerfile.database
index 1eebc26731..585a81a2a7 100644
--- a/docker/Dockerfile.database
+++ b/docker/Dockerfile.database
@@ -26,10 +26,8 @@ RUN apk add --no-cache \
     npm \
     libsndfile
 
-ENV PRISMA_BINARY_CACHE_DIR=/app/.cache/prisma-python/binaries \
-    UV_PROJECT_ENVIRONMENT=/app/.venv \
+ENV UV_PROJECT_ENVIRONMENT=/app/.venv \
     UV_LINK_MODE=copy \
-    XDG_CACHE_HOME=/app/.cache \
     PATH="/app/.venv/bin:${PATH}"
 
 # Copy dependency metadata first for layer caching
@@ -95,18 +93,15 @@ WORKDIR /app
 ENV PATH="/app/.venv/bin:${PATH}"
 
 COPY --from=builder /app /app
+# Prisma binaries live in $HOME/.cache (default prisma-python location),
+# which is /root/.cache here. Copy them from the builder so they survive
+# deployments that volume-mount /app/.cache (e.g. readOnlyRootFilesystem
+# + emptyDir) — otherwise the mount would shadow the baked-in query engine.
+COPY --from=builder /root/.cache /root/.cache
 
 RUN find /app/.venv -type f -path "*/tornado/test/*" -delete && \
     find /app/.venv -type d -path "*/tornado/test" -delete
 
-# Regenerate the Prisma client in the runtime stage so the baked-in
-# BINARY_PATHS resolve to a location outside /app. Users with volume mounts
-# that shadow /app/.cache (e.g. readOnlyRootFilesystem + emptyDir) would
-# otherwise lose access to the pre-downloaded query engine at runtime.
-# Drop the builder's /app/.cache afterwards — it's stale and adds ~800 MB
-# the runtime doesn't use.
-RUN rm -rf /app/.cache && prisma generate --schema=./schema.prisma
-
 EXPOSE 4000/tcp
 
 COPY docker/supervisord.conf /etc/supervisord.conf