How to Deploy OpenGPTs with Docker

I will demonstrate how to quickly deploy an OpenGPTs service using Docker and Docker Compose.

1. Create a file named Dockerfile and fill it with the following content:

FROM node:18 AS frontend
WORKDIR /frontend
COPY ./opengpts/frontend .
RUN npm i
RUN npm run build

FROM python:3.11-slim
WORKDIR /backend
COPY ./opengpts/backend .
RUN rm poetry.lock
RUN pip install .
COPY --from=frontend /frontend/dist ./ui
CMD exec uvicorn app.server:app --host 0.0.0.0 --port $PORT

Do not execute this Dockerfile directly through Docker for now.

2. Create a file named docker-compose.yml and fill it with the following content:

version: "3"
services:
  server:
    build: 
      context: .
    environment:
      - PORT=8000
      - OPENAI_API_KEY=sk-...
    ports:
      - 8000:8000

Replace OPENAI_API_KEY with your own API key.

3. Clone the OpenGPTs project from GitHub:

Use the following command to clone the OpenGPTs project from GitHub:

git clone https://github.com/langchain-ai/opengpts

4. Build the Docker image:

Use the following command to build the Docker image:

docker-compose build

5. Start the service:

Use the following command to start the service:

docker-compose up -d

Now, you can access the deployed service interface through your browser at http://localhost:8000/. If deployed remotely, replace localhost with the corresponding IP address.

For more content, please visit: https://www.freegpttools.org/opengpts