Dockerized openai-realtime-console server not accessible from browser

[gwpl]

Given that this project relies on many dependencies, could you consider providing Dockerfile and run_in_docker.sh , that will allow smooth build and running on different platforms ? (especially Linux, so could be also run_in_docker_for_linux.sh for now?)

[gwpl]

Dockerized openai-realtime-console server not accessible from browser*

---

Description

When running the openai-realtime-console project in a Docker container, the server starts and logs Server listening at http://127.0.0.1:3000. However, it is not accessible from the host browser.

---

Steps to Reproduce

1. Clone the repository:

   git clone https://github.com/openai/openai-realtime-console.git
   cd openai-realtime-console

2. Create a Dockerfile:

   FROM node:18-bullseye
   WORKDIR /usr/src/app
   COPY package*.json ./
   RUN npm install
   COPY . .
   RUN npm run build
   EXPOSE 3000
   CMD ["npm", "start"]

3. Build the Docker image:

   docker build -t openai-realtime-console .

4. Run the Docker container:

   docker run -p 3000:3000 -e OPENAI_API_KEY="$OPENAI_API_KEY" openai-realtime-console

5. Open http://localhost:3000 in a browser.

---

Expected Behavior

The server should be accessible from the browser at http://localhost:3000.

---

Actual Behavior

The server is not accessible from the host browser. Logs show:

2024-12-22 16:55:44.279+0000 - info - Server listening at http://[::1]:3000
2024-12-22 16:55:44.280+0000 - info - Server listening at http://127.0.0.1:3000

Dockerfile_github_only_allows_markdown_attachement.md

Bonus points: if you provide working Dockerfile and README.md instructions helping users to run it safely isolated not only in Docker but also using Kata Containers and/or gVisor.