Deploy

Package your Mountaineer app in a Docker image to deploy it to a container host. The stages below install dependencies, build the frontend, and copy the runtime files into the final image. Docker can cache the dependency stages when only application code changes.

Docker

Exclude local dependencies and generated files from the build context with .dockerignore:

**/node_modules
**/_server
**/_ssr
**/_static
**/_metadata

Image 1: Frontend Dependencies

The first stage uses npm to install frontend dependencies:

FROM node:20-slim as node-dependencies

WORKDIR /usr/src/app

# We only require the dependency definitions
COPY {my_webapp}/views/package.json {my_webapp}/views/package-lock.json ./
RUN npm install

Image 2: Python Builder

This stage installs uv, resolves your Python dependencies into a project virtual environment, and builds the frontend artifacts. Copying pyproject.toml, uv.lock, and README.md before the app source lets Docker cache dependency resolution separately from application code changes.

FROM --platform=linux/amd64 python:3.14-slim-bookworm AS builder

WORKDIR /app

# You only need `nodejs` here if you are using the postcss plugin
RUN apt-get update \
    && apt-get install -y --no-install-recommends nodejs curl \
    && rm -rf /var/lib/apt/lists/*

COPY --from=ghcr.io/astral-sh/uv:latest /uv /uvx /bin/

RUN uv venv /app/.venv
ENV VIRTUAL_ENV=/app/.venv
ENV PATH="/app/.venv/bin:${PATH}"
ENV UV_LINK_MODE=copy

# Copy dependency definitions first for better layer caching
COPY pyproject.toml uv.lock README.md ./

RUN --mount=type=cache,target=/root/.cache/uv \
    uv sync --frozen --no-install-project

# Copy the application code
COPY {my_webapp} ./{my_webapp}
COPY --from=node-dependencies /usr/src/app/node_modules ./{my_webapp}/views/node_modules

# Build the frontend/script files
RUN --mount=type=cache,target=/root/.cache/uv \
    uv run build

Image 3: Final Layer

Copy the Python application, virtual environment, and built frontend into the runtime image. Run it as a non-root user:

FROM --platform=linux/amd64 python:3.14-slim-bookworm AS final

# Create and switch to a new user
RUN useradd --create-home appuser
WORKDIR /app

COPY --from=builder --chown=appuser:appuser /app/.venv /app/.venv
COPY --from=builder --chown=appuser:appuser /app/{my_webapp} ./{my_webapp}

USER appuser

ENV VIRTUAL_ENV=/app/.venv
ENV PATH="/app/.venv/bin:${PATH}"

# Run the application
CMD ["uvicorn", "{my_webapp}.main:app", "--host", "0.0.0.0", "--port", "3000"]

Local Testing

Test the built image locally with docker run. To check production mode outside the container, run:

uv run build
ENVIRONMENT=PRODUCTION uv run uvicorn {my_webapp}.main:app --host localhost --port 5006

This runs with production-minified assets and the same configuration as the production server.

Common Errors

If you see "required file not found" when you try to run this docker image, double check that your venv is pointing to the correct version of Python within the container:

$ ls -ls /venv/bin

4 -rw-r--r-- 1 root root 2209 Mar 29 00:26 activate
4 -rw-r--r-- 1 root root 1476 Mar 29 00:26 activate.csh
4 -rw-r--r-- 1 root root 3039 Mar 29 00:26 activate.fish
4 -rw-r--r-- 1 root root 2724 Mar 29 00:26 activate.nu
4 -rw-r--r-- 1 root root 1650 Mar 29 00:26 activate.ps1
4 -rw-r--r-- 1 root root 1337 Mar 29 00:26 activate_this.py
4 -rwxr-xr-x 1 root root  212 Mar 29 00:27 dotenv
4 -rwxr-xr-x 1 root root  204 Mar 29 00:27 httpx
0 lrwxrwxrwx 1 root root   25 Mar 29 00:26 python -> /usr/local/bin/python3.14
0 lrwxrwxrwx 1 root root    6 Mar 29 00:26 python3 -> python
0 lrwxrwxrwx 1 root root    6 Mar 29 00:26 python3.14 -> python
4 -rwxr-xr-x 1 root root  207 Mar 29 00:27 tqdm
4 -rwxr-xr-x 1 root root  211 Mar 29 00:27 uvicorn
4 -rwxr-xr-x 1 root root  211 Mar 29 00:27 watchfiles
4 -rwxr-xr-x 1 root root  217 Mar 29 00:27 watchmedo

The path python -> /usr/local/bin/python3.14 should be executable and run the interpreter:

$ /venv/bin/python --version

Python 3.14.3