SSH & remote desktop
Connect to the command line with an SSH key. For a graphical desktop, set up VNC and choose a resolution. Access inference ports through an SSH tunnel rather than exposing them directly.
Connection FAQs
Browse by topic
Search first—most questions are already answered
Connecting to a physical node, deploying MLX, or checking cycle pricing? Start with a keyword or category. For questions about your order or instance, sign in to the console and submit a support ticket.
Or choose a topic below.
Jump to a section
Start with your question
These four categories link to the steps and FAQs below. They don’t hide any content, so you can keep reading down the page.
Billing & renewals
Compare daily, weekly, monthly, and quarterly plans. See how the Mac and storage add-ons make up your total, and what to check before your plan expires.
Billing FAQsMLX & AI deployment
Install mlx-lm in an isolated Python environment, start the inference server locally on your physical node, then securely forward its port to your development machine.
Deployment guideAccount & security
View orders, instances, and connection credentials after signing in. Never send your private key by email. For instance-specific issues, submit a support ticket from the console to keep the relevant context together.
Account FAQs
From command line to running service
MLX deployment quick start
Run the commands below on your signed-in Apple Silicon physical node. Run the port-forwarding command on your local computer. First, confirm the instance is ready in the console and keep your SSH connection details handy.
-
01
Create an isolated environment and install mlx-lm
Keep dependencies in a project environment instead of sharing Python packages with existing build tools. After activating it, check that the installation command completes successfully.
python3 -m venv ~/.venvs/mlx source ~/.venvs/mlx/bin/activate python -m pip install --upgrade pip mlx-lm -
02
Listen for inference requests on the node only
This example uses a 7B quantized model. The first load downloads the model files. Available memory also depends on the model, context length, and other running tasks. Keep the listening address set to
127.0.0.1.python -m mlx_lm.server \ --model mlx-community/Qwen2.5-7B-Instruct-4bit \ --host 127.0.0.1 \ --port 8080 -
03
Create an SSH tunnel from your local machine
Set the local environment variables using the connection details shown in the console:
VMR_USERandVMR_HOST. Once the tunnel is up, your local app can connect to127.0.0.1:8080. Traffic is forwarded to the node over SSH. Keep this terminal session open.ssh -N -L 8080:127.0.0.1:8080 \ "${VMR_USER}@${VMR_HOST}"
Check your access boundaries
Don’t change the example’s local listening address to a public-facing address. If several people use the same node, manage SSH public keys separately. For connection details about a specific instance, sign in to the console and submit a support ticket.
Understand your bill
Billing FAQs, by cycle
VMR Lab M4 and VMR Lab M4 Pro are dedicated physical Macs, not virtual machines. Choose a model, billing cycle, and add-ons, then refer to the order generated in the console. Check individual model and add-on prices on the plans page.
How are daily, weekly, monthly, and quarterly plans priced?
Choose a billing cycle to see the corresponding Mac and add-on charges in your order. Each cycle has its own listed price—it isn’t calculated by multiplying the daily rate by the number of days. Before paying, check the cycle, node, add-ons, and total due in USD. For the full price list, seePlans.
Can I change my Mac plan while it’s in use?
The two configurations are M4 / 16GB / 256GB and M4 Pro / 64GB / 2TB. A change may affect your instance and data, so don’t assume your existing environment will transfer unchanged. Check your current order in the console, then submit a support ticket with your target model, current workload, and preferred changeover time. Available options are confirmed case by case.
How are SSD and Thunderbolt 5 add-ons priced?
+1TB SSD, +2TB SSD, and Thunderbolt 5 daisy chaining each have separate daily, weekly, monthly, and quarterly prices. When ordering, check the add-on price for your selected cycle; don’t apply a price from one cycle to another. Thunderbolt 5 daisy chaining is priced per Mac. See the full prices in theplans and add-ons catalog.
Which payment methods and currencies are supported?
Pay with USDT-TRC20 or Visa, Mastercard, and Amex processed by Stripe. All charges are in USD. Check the console for payment options available at checkout. For payment or receipt questions, include your order number in a support ticket—never send your full card details.
What should I check before my plan expires?
Check the order term, renewal options, and backups of important files in the console. Don’t rely on your local calendar to determine order status; use the details and available next steps shown for that order in the console. To change your cycle, share your current instance and preferred cycle in a support ticket so the team can confirm what’s possible.
Command line & graphical desktop
Connection FAQs: choose the right method
Use SSH for commands, file transfers, and port forwarding; use VNC for the macOS graphical desktop. Find connection details in your own instance settings—example addresses aren’t real node addresses.
How do I sign in with an SSH key?
Generate a key pair locally and store it securely. Add the public key to your instance using the process shown in the console. Keep the private key on a device you control. Connect with the username and address in your instance details. If public-key authentication fails, check which private key you’re using, its file permissions, and the public key associated with the instance. Never paste your private key into a support ticket.
How do I fix an unsuitable VNC resolution?
First, make sure the remote desktop session is connected. Then choose a resolution suited to your local display in macOS display settings. Higher resolutions require more bandwidth and scaling. If text looks too small, try lowering the resolution or adjusting the zoom in your local viewer. Get connection credentials for a rented instance from the console.
How do I choose a node in another region?
Choose from Singapore, Tokyo, Seoul, Hong Kong, and the US West Coast. Consider the distance between your team and the node, then how build artifacts or model files will be transferred. Interactive VNC is more sensitive to round-trip latency, while batch jobs, CI, and background inference are usually better placed based on data location and team workflows. Actual performance depends on network routing, so no specific latency can be guaranteed.
What if local port 8080 is already in use?
Change only the local port at the start of the SSH command. For example, use -L 18080:127.0.0.1:8080 as the forwarding option, then connect locally at 127.0.0.1:18080. The service on the node still listens on port 8080. If the tunnel drops, check the SSH session first, then check that the model server is running.
Orders & credentials
Get help with an instance—include the details
Public documentation doesn’t show your instance address or order status. After signing in to the console, check the order number, model, node, and what you were doing when the issue occurred. When submitting a support ticket, include the command needed to reproduce the issue and a redacted error message; this is more useful than saying only “I can’t connect.”
Include in your support ticket
Order number, instance node, commands run, error output, whether the issue is ongoing, and troubleshooting steps already tried. For billing questions, specify the cycle and add-ons to check.
Never include in your support ticket
SSH private keys, full payment card details, or unrelated project secrets. Share redacted logs to describe the issue. To check credentials, first look up the relevant instance in the console.
Technical guides
Explore complete workflows
Links below are populated from published posts on the site blog and shown by publication date. If there are no published posts, no placeholder titles or dates are displayed. For immediate connection and deployment steps, use the quick guides on this page.
Technical articles are being prepared. In the meantime, use the MLX deployment and connection guides on this page, or visit the blog for published posts.
Visit the tech blog
Need more help?
Send your question to our support team
For issues with a rented instance, submit a support ticket from the console so the team can review your order and instance details together. For pre-sales questions about choosing a model, nodes, or team plans, email support@vmrental.comor check the contact page for what to include. Response times vary depending on the question and the checks needed; follow the actual reply in your support ticket.
Ready to go
Choose a physical Mac for your workload
From M4 16GB to M4 Pro 64GB, choose a model based on your memory needs, then check the billing cycle and node. Availability is updated in the console.