Expose Local Apps with Tunnels

Quickly make available applications or tests running on Intel® Tiber™ AI Cloud via a public URL for demos, hackathons, or collaboration. Try the Cloudflare* tool or the Gradio* tool to expose local services to the public.

This tutorial shows how to set up tunnels for various workloads (e.g., web services, web apps, TGI servers, vLLM services). These examples assume your service is running locally on a specific port, e.g., http://localhost:8080.

Prerequisites

  • Linux operating system (OS)

  • Create an instance with Get Started

  • Use One-Click connection for your instance

Optional

  • You may have an image generation service running locally

  • You may need to quickly create a public URL to demo it or test its functionality

Cloudflare TryCloudflare

Cloudflare TryCloudflare lets you quickly expose local services to the Internet with no Cloudflare account required. While intended for testing or temporary use, Cloudflare also offers premium tunnels for production-grade needs.

Benefits of Cloudflare Tool

  • Quick Setup: A single command is enough to expose your service

  • Secure: Uses Cloudflare’s global network, protecting your IP address

  • Temporary Usage: Ideal for demos or quick sharing during development

Set Up TryCloudflare

  1. Access your instance. See Connect to Instance.

  2. Install cloudflared.

    sudo apt update
sudo apt install -y wget
wget https://github.com/cloudflare/cloudflared/releases/latest/download/cloudflared-linux-amd64.deb
sudo dpkg -i cloudflared-linux-amd64.deb

  3. Expose a Local Service. Replace port 8080 with your service port number.

    cloudflared tunnel --url http://localhost:8080

  4. Share the Public URL.

    After running the command, you’ll receive a public URL like:

    https://<random-subdomain>.trycloudflare.com

    Note

    This link is temporary and only works while the cloudflared process is running.

Example Workloads

  • Web Service: Start your web service (e.g., Flask, FastAPI, Node.js) locally on port 8080 and use cloudflared to expose it.

  • TGI or VLLM Server: Run your inference server locally, bound to a port, and use cloudflared to expose it for collaborators or clients to test.

For Production Use

For permanent, production-grade tunnels with custom domains, better reliability, and SLAs, consider upgrading to Cloudflare premium tunnel services.

Gradio Share Feature

Gradio is widely used for creating interfaces for machine learning models, but its share=True option can also be repurposed to expose general-purpose applications.

Benefits of Gradio Tool

  • Ease of Use: Designed for developers, with minimal configuration required

  • Temporary Links: Ideal for demos with links valid for up to 72 hours

  • Integrated with Python: Perfect for exposing Python-based services

Prerequisites

  • pip

  • python3

Set Up Gradio Tunnels

  1. Access your instance. See Connect to Instance.

  2. Install Gradio.

    pip install gradio

  3. Expose a Local Service.

    Create a minimal Gradio app to expose your service (replace 8080 with your service’s port):

    import gradio as gr

def expose_app():
   return "Service running on http://localhost:8080"

demo = gr.Interface(fn=expose_app, inputs=[], outputs="text")
demo.launch(share=True)

  4. Share the Public URL.

    Running the script provides a public URL like:

    https://<random-subdomain>.gradio.live

    Note

    This link is active for 72 hours.

Example Workloads

  • Web Service: Use Gradio to describe and share your web service running on localhost:8080.

  • Web App: Expose your local React, Vue, or other frontend applications by sharing the port they’re hosted on.

  • Inference Servers: Provide lightweight interaction interfaces for machine learning models.

For Production Use

Gradio share links are temporary and not intended for production. For permanent hosting, consider deploying on platforms like Hugging Face Spaces.