diff --git a/.dockerignore b/.dockerignore
new file mode 100644
index 0000000..8b6e390
--- /dev/null
+++ b/.dockerignore
@@ -0,0 +1,4 @@
+build*/
+datasets/
+*.ply
+*.tar.gz
diff --git a/.gitignore b/.gitignore
index 18decc9..6a87295 100644
--- a/.gitignore
+++ b/.gitignore
@@ -20,3 +20,4 @@ large_tool_results/
 
 # Point-cloud exports
 *.ply
+build-docker/
diff --git a/Dockerfile b/Dockerfile
index 1362117..443ec98 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -1,38 +1,39 @@
-# Use Ubuntu 24.04 as base (matching development environment)
+# Build-environment image: contains the toolchain and dependencies only.
+# The source tree is bind-mounted at runtime and compiled inside the container:
+#
+#   docker build -t cloud-point-rpc-dev .
+#   docker run -d --name cprpc-dev --network=host -v "$(pwd)":/app cloud-point-rpc-dev
+#   docker exec -it cprpc-dev meson setup build-docker
+#   docker exec -it cprpc-dev meson compile -C build-docker
+#   docker exec -it cprpc-dev ./build-docker/src/cloud_point_rpc_cli config.yml
+#
+# Use a dedicated build directory (build-docker) so host and container
+# builds never share Meson state.
 FROM ubuntu:24.04
 
-# Avoid interactive prompts during package installation
 ENV DEBIAN_FRONTEND=noninteractive
 
-# Install dependencies
-# - build-essential: Compiler (gcc/g++)
-# - meson/ninja-build: Build system
-# - git: For fetching subprojects
-# - pkg-config, cmake: For dependency resolution
-# - libssl-dev: Often needed for cmake fetches/networking
-RUN apt-get update && apt-get install -y \
+# - build-essential, meson, ninja-build, cmake, pkg-config: toolchain
+# - git: Meson subprojects (asio, json, yaml-cpp, glog, jsonrpccxx, gtest)
+# - libopencv-dev: OpenCV 4 incl. contrib modules (ximgproc for the WLS filter)
+# - libunwind-dev: glog stack traces
+RUN apt-get update && apt-get install -y --no-install-recommends \
     build-essential \
     meson \
     ninja-build \
-    git \
-    pkg-config \
     cmake \
+    pkg-config \
+    git \
     ca-certificates \
     libunwind-dev \
+    libopencv-dev \
     && rm -rf /var/lib/apt/lists/*
 
-# Set working directory
+# The mounted checkout is usually owned by a different uid than root inside
+# the container; git refuses to touch it otherwise (Meson subprojects need git).
+RUN git config --global --add safe.directory '*'
+
 WORKDIR /app
 
-# Copy project files
-COPY . .
-
-# Setup build directory and compile
-# We allow git to fetch subprojects (glog, gtest, asio, etc.)
-RUN meson setup build && \
-    meson compile -C build
-
-# Run the cli by default
-# We assume the config.yaml is in the root /app or we copy it. 
-# The build output is in build/src/cloud_point_rpc_cli
-CMD ["./build/src/cloud_point_rpc_cli", "config.yaml"]
+# Keep the container alive; build and run via `docker exec`.
+CMD ["tail", "-F", "/dev/null"]
diff --git a/README.md b/README.md
index dd025e7..7343af5 100644
--- a/README.md
+++ b/README.md
@@ -20,10 +20,9 @@ Done:
 - [x] Binary PLY export with grid triangulation and `ply_stride` decimation
 - [x] Optional WLS / median disparity post-filters (`wls_filter`, `CpuStereoMatcher::Params`)
 - [x] SCARED benchmark with ground-truth metrics, depth PNGs and `scripts/scared_overview.py`
+- [x] Unity-side C# implementation per [docs/unity-integration.md](docs/unity-integration.md)
 
 To do:
-
-- [ ] Unity-side C# implementation per [docs/unity-integration.md](docs/unity-integration.md)
 - [ ] Remove sliver triangles at the border of the valid region (long spikes are still visible in
       shaded viewers at `ply_stride: 4` because 5 % of local depth allows 3 mm edges)
 - [ ] Visually compare raw vs WLS meshes in a shaded web viewer and decide the visualisation default
@@ -127,27 +126,43 @@ meson test -C build -v
 
 ## Docker
 
-You can build and run the cli using `Docker`.
+The `Dockerfile` builds a **development environment** image only: toolchain
+(GCC, Meson, Ninja, CMake), git for the Meson subprojects and OpenCV 4 with
+contrib modules. Nothing is compiled at image build time. The source tree is
+bind-mounted into the running container and compiled there, so edits on the
+host are picked up immediately and the build artefacts land in your checkout.
 
-### 1. Build Image
+### 1. Build the environment image
 
 ```bash
-docker build -t cloud-point-rpc .
+docker build -t cloud-point-rpc-dev .
 ```
 
-### 2. Run Container
-
-The cli will try to connect to a **running server** on ip and port defined in config.yml file. (defined in `config.yaml` inside the image).
-For simplicity, it's better to use a host network, so you will not have any headache with accessability.
-
-> _Server is not configured to run through container, if you need, contact me_
-
-You also can mount your own `config.yaml` to override the default settings:
+### 2. Start the container with the source mounted
 
 ```bash
-docker run --network=host -it -v $(pwd)/my_config.yaml:/app/config.yaml cloud-point-rpc
+docker run -d --name cprpc-dev --network=host -v "$(pwd)":/app cloud-point-rpc-dev
 ```
 
+The container idles (`tail -F /dev/null`); `--network=host` lets the CLI
+reach a server running on the host and lets the server be reached from Unity.
+
+### 3. Build and run inside the container
+
+```bash
+docker exec -it cprpc-dev meson setup build-docker
+docker exec -it cprpc-dev meson compile -C build-docker
+docker exec -it cprpc-dev meson test -C build-docker
+docker exec -it cprpc-dev ./build-docker/src/cloud_point_rpc_cli config.yml
+```
+
+Use a dedicated build directory such as `build-docker`: Meson stores absolute
+compiler paths, so a build directory configured on the host cannot be reused
+inside the container and vice versa. With rootless Docker the container's
+root maps to your host user, so `build-docker/` stays owned by you; with a
+rootful daemon add `--user "$(id -u):$(id -g)"` to `docker run` to avoid
+root-owned build files.
+
 ## Validation with SCARED Dataset
 
 The `scared_dataset_server` executable lets you validate the stereo point-cloud