Skip to main content

Rivet Server Docker Image

The serve command from the Rivet CLI is available as a Docker image, allowing you to run a Rivet server in a containerized environment.

Quick Start​

# Start a server on port 3000 with a mounted project directory
docker run -p 3000:3000 -v /path/to-project:/project abrennekeironclad/rivet-server:latest

Description​

The Docker image provides the same functionality as the serve command, but packaged in a container for easier deployment and consistent environments. The server runs on port 3000 by default inside the container.

For more information on the serve command, see the Rivet CLI documentation.

Usage​

Basic Usage​

Mount your project directory and expose the server port:

docker run \
  -p 3000:3000 \
  -v /path/to/project:/project \
  abrennekeironclad/rivet-server:latest

With OpenAI Configuration​

To use OpenAI features, provide your API key as an environment variable:

docker run \
  -p 3000:3000 \
  -v /path/to/project:/project \
  -e OPENAI_API_KEY=your-api-key \
  abrennekeironclad/rivet-server:latest

Custom Port​

To use a different port:

docker run \
  -p 8080:3000 \
  -v /path/to/project:/project \
  abrennekeironclad/rivet-server:latest

To Pass Additional Arguments​

To pass additional arguments to the server, append them to the end of the command:

docker run \
  -p 3000:3000 \
  -v /path/to/project:/project \
  abrennekeironclad/rivet-server:latest --dev --allow-specifying-graph-id

Building Custom Images​

Instead of mounting your project at runtime, you can create your own Docker image that includes your Rivet project files. This is useful when you want to:

  • Bundle your project files directly in the image
  • Build a self-contained deployment artifact
  • Set default CLI arguments

Basic Example​

FROM abrennekeironclad/rivet-server:latest

# Copy project files into the image
COPY ./my-project-dir /project

# Optionally set default CLI arguments
CMD ["--dev", "--allow-specifying-graph-id"]

Then, build and run the image:

docker build -t my-rivet-server .
docker run -p 3000:3000 my-rivet-server

Environment Variables​

See the Rivet CLI documentation for a list of environment variables that can be used with the server.

Volume Mounting​

The container expects your Rivet project file to be mounted at /project inside the container. The mounted directory should contain the Rivet project file and any other files needed by the project, such as any .rivet-data file.

Examples​

Using Docker Compose​

version: '3'
services:
  rivet:
    image: abrennekeironclad/rivet-server:latest
    ports:
      - '3000:3000'
    volumes:
      - ./my-project:/project
    environment:
      - OPENAI_API_KEY=${OPENAI_API_KEY}

Using Kubernetes​

You can deploy the Rivet server to Kubernetes using a ConfigMap for your project file and environment variables for configuration.

apiVersion: v1
kind: ConfigMap
metadata:
  name: rivet-project
data:
  project.rivet-project: |
    # Your project file contents here
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: rivet-server
spec:
  replicas: 1
  selector:
    matchLabels:
      app: rivet-server
  template:
    metadata:
      labels:
        app: rivet-server
    spec:
      containers:
        - name: rivet-server
          image: abrennekeironclad/rivet-server:latest
          ports:
            - containerPort: 3000
          volumeMounts:
            - name: project-volume
              mountPath: /project
      volumes:
        - name: project-volume
          configMap:
            name: rivet-project
---
apiVersion: v1
kind: Service
metadata:
  name: rivet-server
spec:
  selector:
    app: rivet-server
  ports:
    - port: 80
      targetPort: 3000
  type: ClusterIP