How to install an AI agent on an OVHcloud VPS

Objective

This guide explains how to deploy an on-premises AI agent on an OVHcloud VPS, without depending on the cloud of external providers. You will use a ready-to-use Docker container containing an open-source AI agent, such as Open Interpreter, GPT4All or Auto-GPT.

Find out how to deploy an AI agent such as Open Interpreter or GPT4All on an OVHcloud VPS.

Requirements

• An OVHcloud VPS (Debian 11 or higher is recommended)
• Administrative (sudo) SSH access to your server
• Python ≥ 3.10 installed on the VPS

Instructions

Update your VPS and install Python

Open a terminal and connect to your VPS with the following command (replacing VPS_IP with the real IP):

ssh <user>@VPS_IP

Update the packages:

sudo apt update && sudo apt upgrade -y

Verify that Python ≥ 3.10 is installed:

python3 --version

If necessary, install Python 3.10+ and pip:

sudo apt install -y python3 python3-pip python3-venv

Create a virtual environment and install Open Interpreter

Recent environments limit the use of pip globally. It is recommended that you create and use a virtual environment:

python3 -m venv ~/venv-openinterpreter
source ~/venv-openinterpreter/bin/activate

Install Open Interpreter:

pip install --upgrade pip
pip install open-interpreter

Configure an AI template for the agent (local or remote)

Option 1 - Use OpenAI (GPT-4o, GPT-3.5, etc.)

You must have an OpenAI API key. Add it when you launch it for the first time, or set it beforehand:

export OPENAI_API_KEY="API_KEY"
interpret

Then follow the instructions.

Option 2 - Use a local model (via Ollama)

If you do not want to use OpenAI, run a local model using Ollama:

interpreter --local

This will give you several options (Ollama, LM Studio, etc.). We recommend using Ollama, as it is easy to install and compatible with several modern models such as llama3 or mistral.

Install Ollama:

sudo apt install curl -y
curl -fsSL https://ollama.com/install.sh | sh
exec $SHELL

Load the model:

ollama pull mistral

Then try again:

interpreter --local

Common errors

Model requires more system memory (6 GiB) than is available (1 GiB)

This error indicates that your VPS does not have enough RAM. Here are your options:

• Choose a new VPS with at least 8 GB RAM.
• Use the OpenAI API (see Option 1 above).
• Use a lighter model, like mistral.

Test your AI Agent

Test the following use cases:

What is in the /tmp folder?

Sample response:

Write a Python script that lists the files in the current folder.

The agent interprets your request, generates code, and executes it locally.

Sample response:

Conclusion

With this guide, you have installed an AI agent on your OVHcloud VPS, capable of executing commands from simple natural language instructions. Whether you have opted for a remote model like GPT-4 via OpenAI or a local model like Mistral via Ollama, you now have an intelligent assistant directly in your terminal, without dependency on a web interface or a third-party service.

Go further

For specialized services (SEO, development, etc.), contact OVHcloud partners

Join our community of users.