Key highlights
- Learn how to configure a VPS to keep Hermes Agent running 24/7 without manual restarts.
- Understand the Hermes Agent architecture and which AI model works best for continuous autonomous operation.
- Explore the full Hermes Agent API setup process, from GitHub installation to connecting your OpenAI credentials.
- Know which common errors affect always-on AI agent servers and how to resolve them quickly.
- Uncover how auto-restart scripts and scheduled tasks keep your self-hosted agent online through crashes and reboots.
A typical Hermes setup works fine on a laptop until you need it to run continuously. For example, if your agent checks API uptime every 6 hours or sends alerts to Slack, it stops the moment your system shuts down or disconnects. That limitation makes local setups unreliable for real automation.
Running Hermes Agent on a VPS solves this by turning it into a persistent, always-on system. You can keep workflows active, schedule recurring jobs and access your agent remotely without depending on your machine. It also gives you better stability, control, and scalability.
In this guide, you’ll learn how to install Hermes Agent, complete the API setup, configure the gateway, and run it as a reliable 24/7 AI agent server.
Why run Hermes Agent on a VPS instead of a laptop?
Hermes Agent can run locally for testing, but a VPS gives it the stable runtime needed for continuous automation. Instead of relying on your laptop’s power state, network connection or local environment, Hermes can keep running on a remote server with predictable resources and full configuration control.
This setup is especially useful when you want to schedule recurring tasks, keep gateway workflows active, access the agent remotely or run long-lived services without interruption. Once that foundation is clear, the next step is choosing the right VPS resources for your workload.
Recommended VPS specs for Hermes agent hosting
Hermes Agent itself is lightweight, but your actual resource needs depend on what the agent is doing. A simple CLI setup requires minimal resources, while continuous automation, browser tasks, or frequent cron jobs will need more CPU and memory. The goal is to start small and scale based on workload.
Suggested VPS specs
| Use case | Recommended specs |
| Basic CLI testing | 1–2 vCPU, 2 GB RAM |
| Hermes Agent 24/7 uptime with light cron jobs | 2 vCPU, 4 GB RAM |
| Automation workflows, gateway use or browser tasks | 4 vCPU, 8 GB RAM |
| Local model inference | Higher-RAM VPS, GPU server or external inference endpoint |
If you’re using hosted model providers like OpenAI or OpenRouter, Nous Portal or another OpenAI-compatible endpoint, your VPS mainly handles the Hermes Agent server, gateway, memory shortage and scheduled tasks. You don’t need a GPU unless you plan to run models locally.
With your VPS properly sized for your workload, you’re ready to move from planning to execution and set up a fully functional, always-on agent.
How to run Hermes agent 24/7 on a VPS
Once your VPS is ready and you understand the requirements, the next step is to move from a basic setup to a fully functional, always-on Hermes Agent server. Before you start running commands, here is the full setup path at a glance.
Quick steps to run Hermes Agent 24/7 on a VPS
| Step | What you do | Outcome |
| 1. Prepare the server | Connect with SSH, update packages and install Git | Clean VPS environment |
| 2. Install Hermes Agent | Run the official installer | Hermes CLI is available |
| 3. Configure the model provider | Run setup and add your API key | Hermes can process tasks |
| 4. Test the setup | Start Hermes and run diagnostics | Confirms the install works |
| 5. Set up the gateway | Configure and test the gateway | Enables always-on workflows |
| 6. Run as a system service | Install and start the gateway service | Hermes keeps running after logout or reboot |
| 7. Add cron jobs | Create scheduled tasks | Automates recurring workflows |
| 8. Secure the server | Harden SSH, firewall and API key handling | Reduces operational risk |
| 9. Troubleshoot | Check service status, logs and diagnostics | Keeps the setup reliable |
Once you understand the flow, follow the steps below to install Hermes Agent, configure the gateway and keep it running continuously on your VPS.
Step 1: Server preparation
Before you install Hermes Agent, your VPS needs to be properly configured for a stable, long-running setup. Since this will become your Hermes Agent server, even small misconfigurations (outdated packages, missing dependencies, wrong user access) can cause failures later when you try to run Hermes Agent 24/7.
Connect to your VPS
Start by accessing your server using SSH. This gives you remote control over your environment:
ssh username@your-server-ip
If you’re using a cloud provider, make sure your firewall allows SSH access (port 22) and you’re connecting with the correct user.
Update your system
Outdated system packages can break installations or cause compatibility issues with dependencies. After connecting to your VPS through SSH, update the server’s package list and installed packages: Run:
sudo apt update && sudo apt upgrade -y
This ensures your Linux environment is secure, stable, and ready for the Hermes Agent install.
Install Git (Required for Hermes agent install)
Hermes uses Git to fetch and install its components. If Git is missing, the installer won’t work. After updating your VPS, install Git on the server:
sudo apt install git -y
Important: Hermes’ installer automatically handles Python, Node.js, and other dependencies, so Git is typically the only manual requirement before setup.
Step 2: Install Hermes agent on your VPS
Now that your server is ready, the next step is to install Hermes Agent. The official installer is designed to simplify the process, so you don’t have to manually configure dependencies or environments.
Run the official Install command
Use the one-line installer to download and set up Hermes:
curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash
This command will:
- Clone the Hermes Agent repository from GitHub
- Set up a virtual environment
- Install required dependencies (Python, Node.js, etc.)
- Register the global hermes CLI command
Reload your shell
After installation, reload your shell so the Hermes CLI becomes available:
source ~/.bashrc
Or, if you use Zsh:
source ~/.zshrc
Start Hermes agent
Launch the CLI to verify installation:
hermes
If everything is set up correctly, Hermes will start and prompt you for initial configuration.
Tip: This installation method works across Linux environments and handles most setup steps automatically, reducing manual errors.
At this point, Hermes Agent is installed on your VPS, but it’s not yet configured or running 24/7.
Step 3: Configure Hermes agent API setup (Model provider)
After installation, Hermes Agent needs a model provider to function. This step connects your agent to an LLM (like OpenAI or OpenRouter) so it can process tasks, generate responses, and run workflows.
Run the setup wizard
The easiest way to configure everything is:
hermes setup
This guided flow helps you:
- Select a model provider
- Add your API key
- Configure default settings
Manually select or change model provider
You can also configure or switch providers anytime using:
hermes model
Add your API key
Depending on the provider you choose, you’ll need to supply an API key. This is typically stored securely in your Hermes configuration or environment variables.
Best practice:
- Never hardcode API keys in prompts
- Store them in .env or Hermes config files
- Rotate keys if exposed
What’s the best model for Hermes agent?
There’s no single “best” model, it depends on your use case:
- Automation & workflows → fast, cost-efficient models
- Coding tasks → higher reasoning models
- General tasks → balanced models
Start with a reliable hosted provider, then optimize based on performance and cost.
At this point, your Hermes Agent API setup is complete. The agent now has the intelligence layer it needs to operate.
Step 4: Test your Hermes agent setup
Before turning Hermes into a 24/7 service, you need to confirm that everything is working correctly. Skipping this step is a common mistake and can lead to silent failures later when the agent runs in the background.
Start Hermes CLI
Run:
hermes
Once it starts, send a simple test prompt like:
- What model are you using?
- Can you access the current working directory?
This verifies that:
- The model provider is configured correctly
- API keys are working
- The agent can execute basic tasks
Run diagnostics
Use the built-in diagnostic tool:
hermes doctor
This command checks for:
- Missing dependencies
- Configuration issues
- API setup problems
Why this step matters
Hermes Agent runs as a background service in later steps. If the base CLI setup is broken, the gateway and cron jobs won’t work properly and debugging becomes harder.
Only move forward once:
- CLI works without errors
- Responses are generated correctly
- hermes doctor shows no critical issues
Now your Hermes Agent server is fully functional at the CLI level.
Step 5: Set up Hermes gateway (Core of 24/7 Uptime)
At this stage, Hermes works in the CLI, but it’s not always running. To enable continuous execution, messaging, and scheduled jobs, you need to set up the Hermes gateway. This is the core component that turns your setup into a true 24/7 AI agent server.
The gateway runs as a long-lived process that:
- Receives and routes messages
- Maintains session context
- Executes scheduled (cron) jobs every ~60 seconds
Configure the gateway
Run the setup command:
hermes gateway setup
This will guide you through configuring messaging platforms, routing behavior, and basic settings.
Test the gateway in foreground mode
Before running it as a service, start it manually:
hermes gateway
This lets you:
- Verify that the gateway starts without errors
- Confirm messages and tasks are processed correctly
- Catch configuration issues early
What to check
- No startup errors in logs
- Agent responds correctly through the gateway
- Cron scheduler initializes properly
Important: The gateway is what enables Hermes Agent to run continuously. Without it, cron jobs won’t execute and your agent won’t function as a background system.
Now your Hermes Agent architecture is in place, but it still needs to be made persistent.
Step 6: Run Hermes agent 24/7 using system service
Now comes the most important step. To keep Hermes running continuously, even after a logout or server reboot, you need to install the gateway as a system service. This is what enables true Hermes Agent 24/7 uptime on a VPS.
Install Hermes gateway as a system service
Run:
sudo hermes gateway install –system
This registers Hermes with your system’s service manager (systemd), allowing it to run in the background.
Start the service
sudo hermes gateway start –system
Check service status
sudo hermes gateway status –system
You should see the service running without errors.
View logs for debugging
journalctl -u hermes-gateway -f
This helps you monitor real-time logs and catch issues early.
Why system service matters
- Runs Hermes automatically after server reboot
- Keeps the agent active after SSH logout
- Eliminates dependency on your session
- Ensures consistent uptime for automation and cron jobs
Alternative: User service (Not recommended for VPS)
You can run Hermes as a user-level service:
hermes gateway install
hermes gateway start
But this requires additional configuration (loginctl enable-linger) and is less reliable for headless environments.
For a production-ready Hermes Agent server, always use the system service setup.
At this point, your Hermes Agent is running 24/7 on your VPS.
Step 7: Enable scheduled jobs (Hermes cron)
Now that your Hermes Agent is running continuously, you can automate tasks using built-in cron functionality. This is what turns your setup from a passive system into an active 24/7 AI agent server.
Hermes cron is managed by the gateway and runs on a loop, checking for scheduled jobs roughly every 60 seconds.
How Hermes cron works
- Jobs are stored in: ~/.hermes/cron/jobs.json
- Output is saved in: ~/.hermes/cron/output/
- Each job runs in a fresh agent session
- Results are logged and optionally sent via integrations
Useful commands
hermes cron list
hermes cron status
hermes cron pause
hermes cron resume
hermes cron run
hermes cron remove
Example use case
You can create a job like:
- Check your app every 6 hours
- If status is OK → return [SILENT]
- If not → send alert with details
This allows you to automate monitoring, reporting, and workflows without manual intervention.
Key Things to remember
- Cron jobs only run when the gateway is active
- Timing depends on your server timezone
- Jobs support intervals, cron expressions, and timestamps
If your jobs don’t run, always check whether the gateway service is running.
At this point, your Hermes Agent is not just running, it’s actively executing automated workflows.
Step 8: Security checklist for your Hermes agent server
Once your Hermes Agent is running 24/7, security becomes critical. Since your agent can access APIs, run commands, and automate workflows, a poorly secured server can expose sensitive data or allow unauthorized access.
Use SSH keys instead of passwords
Disable password-based login and use SSH keys for secure access. This significantly reduces the risk of brute-force attacks.
Disable root login
Edit your SSH config (/etc/ssh/sshd_config) and set:
PermitRootLogin no
Always use a non-root user with sudo privileges to run Hermes.
Configure firewall rules
Limit access to only required ports:
sudo ufw allow 22
sudo ufw enable
You can further restrict access to specific IPs for added security.
Secure API keys
- Store keys in environment files or Hermes config
- Never expose keys in prompts or logs
- Rotate keys if compromised
Monitor logs regularly
Check service logs to detect unusual behavior:
journalctl -u hermes-gateway -f
Back up Hermes data
Backup your .hermes directory regularly. It may contain:
- Agent memory
- Skills
- Cron jobs and outputs
Additional best practices
- Keep your system updated
- Review tools and permissions before enabling them
- Restrict access to messaging integrations (Telegram, Slack, etc.)
A secure setup ensures your Hermes Agent server runs reliably without exposing your workflows or credentials.
Step 9: Troubleshoot common Hermes agent issues
Even with a correct setup, issues can happen, especially when running Hermes as a background service. The key is knowing where to look and how to fix problems quickly.
Common issues and fixes
| Problem | Likely cause | Fix |
| hermes: command not found | Shell not reloaded or PATH issue | Run source ~/.bashrc or restart session |
| API key error | Model provider not configured | Run hermes setup or hermes model |
| Cron jobs not running | Gateway not active | Start service using system service |
| Jobs running at wrong time | Server timezone mismatch | Check with date and adjust timezone |
| Gateway stops after logout | Using user service | Switch to system service setup |
| Messaging not working | Misconfigured integration | Re-run hermes gateway setup |
| Missing dependencies | Installation issue | Run hermes doctor |
Check service status
If something breaks, start by checking whether the gateway is running:
sudo hermes gateway status –system
View logs for errors
Logs are the fastest way to identify issues:
journalctl -u hermes-gateway -f
Run diagnostics
Use:
hermes doctor
This helps detect:
- Missing tools
- Configuration errors
- API setup problems
Pro tips
- Always test changes in CLI before applying to service
- Restart the service after major changes
- Verify API keys and provider settings first (most common failure point)
At this stage, you have a fully functional Hermes Agent running 24/7 with automation, scheduling, and debugging in place.
What you can do next with your Hermes agent server
Now that your Hermes Agent is running 24/7 on a VPS, you can move beyond setup and start building real automation. Use it to monitor APIs, run scheduled reports, or trigger alerts through Slack, Telegram, or Discord without relying on your local system. As your workflows grow, optimize performance by choosing the right model, adjusting cron frequency, and monitoring logs to prevent silent failures.
You can also expand your setup by connecting external APIs, adding integrations, or creating multi-step automation pipelines. Running Hermes on a VPS gives you the flexibility to scale, maintain uptime, and build a fully autonomous system that works continuously in the background.
As your usage scales, choosing the right infrastructure becomes just as important as the setup itself.
How to choose the right VPS for Hermes Agent hosting
Choosing the right VPS for Hermes Agent hosting starts with one question: what will your agent actually do?
Hermes Agent itself does not need heavy infrastructure for basic CLI use. But once you start running the gateway continuously, adding scheduled jobs, connecting messaging integrations or using browser-based workflows, your VPS needs enough resources to stay stable without slowing down or dropping background tasks.
Look for a VPS that gives you five things:
1. Always-on availability
Hermes Agent becomes more useful when it can keep running after your laptop shuts down. Choose a VPS built for persistent uptime, so the Hermes gateway can stay active, scheduled jobs can keep running and your agent remains reachable when you are offline.
This is especially important if you are using Hermes for API monitoring, recurring reports, Slack alerts or long-running automation workflows.
2. Full root access
A Hermes Agent VPS should give you control over the server environment. You may need to install packages, manage system services, configure environment variables, review logs or adjust gateway settings.
Root access gives you the flexibility to run Hermes as a persistent service, tune dependencies and troubleshoot issues without being limited by a locked-down hosting environment.
3. Dedicated CPU, RAM and fast storage
For light use, Hermes can run on modest VPS resources. But gateway workflows, cron jobs and browser automation can become resource-intensive over time.
Choose a VPS with dedicated resources and fast NVMe storage, especially if you plan to run multiple recurring jobs, connect several integrations or keep Hermes active for production-like workflows.
A practical starting point is:
| Hermes Agent use case | Suggested VPS profile |
| Basic testing | 1–2 vCPU, 2 GB RAM |
| 24/7 gateway with light cron jobs | 2 vCPU, 4 GB RAM |
| Browser tasks or heavier automation | 4 vCPU, 8 GB RAM |
| Local model inference | Higher-RAM VPS, GPU server or external inference endpoint |
If you use hosted model providers like OpenAI, OpenRouter or another compatible endpoint, your VPS mainly runs Hermes Agent, the gateway, memory storage and scheduled tasks. You do not need a GPU unless you plan to run models locally.
4. Secure self-managed control
Because Hermes Agent can access APIs, execute tools and automate workflows, your VPS should let you apply proper security controls. Look for SSH access, firewall configuration, non-root user setup, API key management and regular backup options.
A self-managed VPS is often the better fit for Hermes Agent because it gives technical users full control over the runtime environment while keeping the agent isolated from their local machine.
5. Room to scale as workflows grow
Your first Hermes setup may start with one scheduled task, but agentic workflows often expand quickly. You may add more cron jobs, more integrations, more tool use or heavier automation.
Choose a VPS that can scale CPU, RAM and storage without forcing you to rebuild your setup from scratch.
Where Bluehost fits
Bluehost Hermes Agent VPS hosting fits this use case when you want a persistent, self-managed environment for running Hermes beyond a local CLI setup. With root access, dedicated VPS resources, NVMe storage and an always-on server environment, Bluehost gives you the foundation needed to run the Hermes gateway, scheduled jobs and long-running agent workflows with more control.
It is especially useful if your goal is to move from local experimentation to a stable remote Hermes Agent server. You can install Hermes, connect your preferred model provider, configure the gateway and keep your agent running continuously without depending on your personal machine.
The key is to choose your VPS based on workload. Start with enough resources for the gateway and light automation, then scale as your Hermes Agent setup grows into more advanced workflows.
Final thoughts
Running Hermes Agent on a VPS turns it from a local experiment into a persistent AI workspace. With the right server specs, model provider, gateway setup, system service, cron jobs and security checks, Hermes can keep working even when your laptop is offline.
Bluehost Hermes Agent VPS hosting gives that setup the always-on foundation, root access and resource control needed for long-running agent workflows.
The bigger question is not just whether Hermes can run 24/7. It is what you will trust your agent to handle once it can.
Frequently asked questions
Can Hermes Agent run 24/7 on a VPS?
Yes. Hermes Agent can run 24/7 on a VPS when you install it on a Linux server, configure a model provider, set up the Hermes gateway and run the gateway as a system service. This keeps the agent active after SSH logout or server reboot.
Do I need a GPU to host Hermes Agent on a VPS?
No, you do not need a GPU if you use hosted model providers like OpenAI, OpenRouter or another compatible endpoint. Your VPS mainly runs Hermes Agent, the gateway, memory storage and scheduled jobs. A GPU is only needed if you plan to run local model inference on the same server.
What VPS specs are best for Hermes Agent hosting?
For light 24/7 use, 2 vCPU and 4 GB RAM is a practical starting point. For browser automation, heavier tools or frequent cron jobs, 4 vCPU and 8 GB RAM gives more stability. Basic CLI testing can run on smaller resources.
How do I keep Hermes Agent running after SSH logout?
Run the Hermes gateway as a system service. A system service keeps Hermes running in the background, even after you close your SSH session or reboot the VPS.
What is the Hermes gateway used for?
The Hermes gateway is the always-on layer that supports messaging, session handling and scheduled jobs. The CLI is useful for direct interaction, but the gateway is what turns Hermes into a persistent 24/7 AI agent server.
Do Hermes cron jobs work without the gateway?
No. Hermes cron jobs depend on the gateway or a running service process. If the gateway is not active, scheduled jobs will not run.
Can I connect my 24/7 Hermes Agent VPS instance to Slack, Telegram or Discord?
Yes. You can connect Hermes Agent to messaging platforms through the Hermes gateway and supported integrations. This lets your agent send alerts, respond to messages and support real-time workflows from your VPS-hosted setup.
Can Hermes Agent be configured to use OpenRouter as a model provider?
Yes. You can configure OpenRouter during the Hermes Agent API setup if your Hermes version supports OpenAI-compatible endpoints. Add your OpenRouter base URL and API key, then test the setup before running the gateway as a service.
Is Hermes Agent better on a VPS than a laptop?
A laptop is fine for testing, but a VPS is better for continuous automation. A VPS keeps Hermes Agent online, supports scheduled workflows, allows remote access and avoids interruptions caused by sleep mode, shutdowns or unstable local networks.
How do I troubleshoot Hermes Agent if it stops running?
Start by checking the gateway service status, reviewing logs and running hermes doctor. Common issues include API key errors, gateway misconfiguration, cron timing problems, missing dependencies or a stopped system service.
Write A Comment