Most AI tools today are still glorified chatbots.
You ask a question.
You get an answer.
And by tomorrow… it forgets you even exist.
But what if you could build an AI assistant that:
🧠 Remembers your preferences
📈 Learns from every interaction
🌐 Researches the web in real time
📱 Runs directly on your phone
⏰ Automates tasks while you sleep
That’s exactly what I built using Hermes, one of the fastest-rising AI agent frameworks right now and increasingly seen as a serious competitor in the personal AI space.
🔥 Why Hermes Is Blowing Up
Hermes isn’t just another chatbot framework.
It introduces something much more powerful:
🧠 Long-Term Memory
Your assistant remembers:
- Your goals
- Your recurring tasks
- Your preferences
- Your habits
🔄 Self-Reflection
After every task, Hermes can evaluate:
- What worked
- What failed
- What should improve next time
This means your AI literally gets smarter the more you use it.
🤖 Real Automation
You can use Hermes for:
- Daily AI news briefings
- Travel planning
- Productivity reminders
- Research tasks
- Personal scheduling
- Coding assistant
- Messaging integrations like Telegram,Discord, etc
🛠️ What You’ll Build in This Tutorial
In this beginner-friendly walkthrough, I cover:
✅ Installing Hermes from scratch
✅ Configuring free Google AI Studio API access
✅ Adding Tavily for live web search
✅ Connecting your AI assistant to Telegram
✅ Creating a true phone-based personal assistant
✅ Security best practices
🎥 Full video walkthrough
⚙️ Installation: Build Your First Hermes AI Assistant
Getting started with Hermes is surprisingly simple — even if you’re completely new to AI agents.
Within minutes, you’ll have your own self-improving assistant running locally.
🚀 Step 1: Run the Hermes Installer
Simply run:
curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash
What this does:
✅ Downloads Hermes
✅ Installs required dependencies
✅ Configures your environment
✅ Guides you through setup interactively
💡Pro Tip: macOS and Linux users can install directly. Windows users should first install WSL2 for the smoothest experience.
🤖 Step 2: Choose Your Model Provider
Hermes supports multiple providers, but for beginners, Google AI Studio (Gemini) is one of the best choices because:
💸 Generous free tier
⚡ Fast setup
🛠️ Beginner-friendly
🚀 Great for experimentation
During setup, select:
(•) Google AI Studio (Gemini models — native Gemini API)
🔑 Step 3: Get Your Google AI Studio API Key
To power Hermes, you’ll need an API key.
Steps:
- Go to Google AI Studio
- Navigate to Projects
- Click Create New Project
- Go to API Keys
- Select your project
- Click Create API Key
- Copy your key
- Paste it into the Hermes installer when prompted
🔒 Security Tip:
Treat API keys like passwords — never share them publicly.
📱 Step 4: Skip Messaging Setup (For Now)
During initial installation, Hermes will ask about messaging integrations.
👉 Skip this step temporarily.
We’ll configure Telegram later to transform Hermes into a true mobile personal assistant.
💬 Step 5: Launch Hermes Chat
When prompted, enter: y
Hermes will launch its built-in TUI (Terminal User Interface).
You now have:
✅ Your first AI agent
✅ Local chat interface
✅ Long-term memory capabilities
✅ Self-improvement foundation
🎉 Congratulations — your personal AI assistant is officially alive.
🔍 Configure Tavily for Real-Time Web Search
Without web access, your AI assistant is limited.
Adding Tavily gives Hermes the ability to:
🌐 Search the internet
📰 Gather live information
📊 Perform research tasks
✈️ Plan trips using current data
🛠️ Tavily Setup Steps
- Create a Tavily account
Visit Tavily and sign up.
- Generate your API key
Create a new key from your dashboard.
- Add it to Hermes
hermes config set TAVILY_API_KEY <YOUR_TAVILY_API_KEY>
🎯 Result:
Your Hermes assistant can now perform real-time web searches and research autonomously.
📲 Connect Hermes to Telegram (Your AI in Your Pocket)
Right now, Hermes only runs on your computer.
Connecting Telegram transforms it into a true personal assistant you can message from anywhere.
🤖 Step 1: Create a Telegram Bot
Instructions:
- Open Telegram
- Search for
@BotFather - Type:
/newbot - Choose a bot name
- Choose a unique username ending in:
_bot - Copy your bot token
🔑 You’ll need this token to connect Hermes.
Step 2: Connect Telegram to Hermes
- Run the following command:
hermes setup gateway
Then:
✅ Select Telegram
✅ Paste your Bot Token
✅ Add your Telegram numeric user ID
✅ Set your account as the home channel
✅ Start the gateway service
📍 How to Find Your Telegram User ID
Search Telegram for: @userinfobot
It will instantly provide your numeric user ID.
🎉 Final Result
Your Hermes AI Assistant is now:
📱 Available on your phone
💬 Reachable through Telegram
🧠 Memory-enabled
🌐 Web-connected
📈 Self-improving
You’ve officially built a beginner-friendly AI agent that can evolve into a true personal productivity system.
🌍 Real-World Examples
Once set up, your AI assistant can:
✈️ Personalized Travel Planning
Ask it to plan a trip to Japan… then later France…
And it remembers:
- Your budget
- Your interests
- Your travel style
📰 Daily Briefings
Example:
“Every morning at 8 AM, send me the top 3 AI news stories.”
Your assistant works while you sleep.
📅 Weekly Productivity Systems
Example:
“Every Friday at 4 PM, remind me to review my week.”
⚠️ Important Security Tips
Hermes is powerful — but with power comes responsibility.
Best Practices:
🔒 Avoid running on your primary machine
🖥️ Use a VM, remote server, or sandbox
🔑 Protect API keys like passwords
⚡ Be cautious with community-built skills
🧠 Why This Is Bigger Than Just One Tutorial
We’re watching AI evolve from:
Old Model:
Chatbot → Answer → Forget
New Model:
Assistant → Learn → Remember → Improve
This is a massive shift.
We’re entering the era of persistent personal AI.
💬 Final Thoughts
Hermes shows where AI is truly heading:
✨ Personalized
✨ Persistent
✨ Autonomous
✨ Self-improving
For beginners, this may be one of the most practical AI agent projects you can build right now.