Unleashing AI Agents
At Aspyn Information Services, we are always exploring ways to streamline IT operations and boost productivity. Recently, we’ve been leveraging the Hermes Agent—a powerful, open-source AI agent framework that runs directly in your terminal, providing persistent context and autonomous system control (It gets to know you and fully customizable!).
One of the most effective ways to get started is by pairing Hermes with a free Gemini API key from Google. AI Tokens can get expensive, especially when you are experimenting. This guide will walk you through setting it up so you can start automating your own IT workflows without breaking the bank!
Why Hermes Agent?
Unlike standard chatbot interfaces, Hermes is designed for deep integration with your environment. It:
* Maintains persistent memory: It remembers your preferences, environment details, and past successes.
* Learns as it works: It captures reusable procedures as “skills,” allowing it to get better at your specific IT tasks over time.
* Works everywhere: Whether you are using a Linux server, a Windows workstation, or a messaging platform, Hermes provides a consistent interface to interact with your systems.
Step 1: Obtain Your Free Gemini API Key
Google offers a generous free tier for its Gemini API, making it an ideal choice for testing and development.
1. Navigate to Google AI Studio (https://aistudio.google.com/).
2. Log in with your Google account.
3. Click the “Get API key” button on the left sidebar.
4. Create a new API key in a new or existing Google Cloud project. (You will likely need to create a new Google Cloud Project – name it something recognizable like “Hermes Agent:
5. Securely copy your key. You will need this for the next step.
Step 2: Configure Hermes Agent
Once you have your API key, configuring Hermes is straightforward. Hermes Agent allows you to manage models and providers easily through its command-line interface.
If You Haven’t Installed Hermes Yet
Run the following command in your terminal:
Install hermes using curl from command line
For our tests, we have been working with Hermes running in an Ubuntu desktop VM. We will cover the install and configuration of Hermes specific to Ubuntu in a future post. (Hermes appears to be “out of the box” tuned for running on a Mac).
You will be prompted to select a variety of “skills” for Hermes to use. To keep things simple, stick with the basics: file manipulation, console/shell, web browser, email support etc. The skills library is large and it has a great ecosystem for custom plugin/skill creation.
At some point in the install you will be prompted to choose a model. These settings can be brought up and configured at any time using the command “hermes model”.
You will be presented with a list of common models as well as the option to define a “custom” model (which we would use to connect to our local hosted QWEN model utilizing a different Hermes profile – will cover in another post).
Model Select
Select Google Gemini and the API option:
Choose Google AI Studio
You will be asked to paste in your key (avoid extra whitespace!). If you mess it up, you can always run the config again!
The system will ask you for the endpoint – just select the default.
The guided walkthrough is the easiest approach; alternatively, you can manually set it in your configuration:
hermes config set model.provider gemini
hermes config set model.api_key YOUR_API_KEY_HERE
To ensure everything is working correctly, run the built-in health check:
hermes doctor
Start Automating!
With Gemini configured, you’re ready to go. You can start a chat session by typing `hermes` in your terminal or use single-query mode for quick tasks:
hermes chat -q “List the files in my current directory and summarize their contents.”
At Aspyn Information Services, we find that the true power of Hermes lies in its ability to build and refine custom skills tailored to your business needs. If you find yourself repeatedly running the same routines, diagnostic commands or deployment scripts, let Hermes turn that process into a reusable skill. Have a report that you need to build, often based on multiple data sources? Hermes can streamline this for you. Do you have a flow that starts with an email and proceeds along an assembly line of data extraction and inputs elsewhere? Hermes might be the answer. The options are endless, but the most effective solutions are found to automate repetitive and straightforward tasks that require just a little bit of inference or context (otherwise it is often better to just script it directly and avoid the LLM overhead).
Need help integrating AI agents into your business operations? Contact us at Aspyn Information Services, and let’s discuss how we can help you stay ahead.
