Docker Service Image

This page covers the standalone Docker image for the NeMo Retriever service. For production-scale service and NIM deployment, use the Helm chart in helm/README.md and the published NeMo Retriever Library install procedures.

Build The Service Image

Run from the repository root:

docker build \
  -f Dockerfile \
  --target service \
  -t nemo-retriever-service:dev \
  .

The service target installs nemo_retriever[service], copies the packaged retriever-service.yaml, and starts the service with:

retriever service start --config /etc/nemo-retriever/retriever-service.yaml

Run The Service Container

For remote NVIDIA-hosted NIM endpoints, pass the API key into the container and publish the service port:

docker run --rm \
  -p 7670:7670 \
  -e NVIDIA_API_KEY="${NVIDIA_API_KEY}" \
  nemo-retriever-service:dev

Open http://localhost:7670/docs for the OpenAPI UI, or check health with:

curl -fsSL http://localhost:7670/v1/health

Configure The Service

The image reads /etc/nemo-retriever/retriever-service.yaml by default. To run with a custom config, mount it and pass the path through the service CLI:

docker run --rm \
  -p 7670:7670 \
  -e NVIDIA_API_KEY="${NVIDIA_API_KEY}" \
  -v "$PWD/my-retriever-service.yaml:/etc/nemo-retriever/retriever-service.yaml:ro" \
  nemo-retriever-service:dev \
  retriever service start --config /etc/nemo-retriever/retriever-service.yaml

Use Kubernetes Secrets, Helm values, or container environment variables for credentials. Do not bake API keys into derived images.

Audio And Video

The service image omits ffmpeg and ffprobe by default. For audio or video extraction on a development machine with package-repository network access, set INSTALL_FFMPEG=true:

docker run --rm \
  -p 7670:7670 \
  -e NVIDIA_API_KEY="${NVIDIA_API_KEY}" \
  -e INSTALL_FFMPEG=true \
  nemo-retriever-service:dev

For restricted or air-gapped environments, build a derived image that includes ffmpeg and ffprobe, then set the Helm service.image.* values or run that derived image directly.