Nothing found with "” matching help entry.
Ordering and Delivery
How to place an order on the official website? How long does it take to deliver?
existOrder onlineOn the page, select the machine model, lease period and data center, and complete the payment intention and information filling. The resource isPhysical machine reservation mode, typical delivery is completed within a few hours to 1 working day after order confirmation (affected by inventory and machine room shelving batches, the console will display the estimated readiness time).
Do I need to prepare an SSH public key when placing an order?
Strongly recommendedUpload when placing an order or initializing
SSH Public Key. Password login is turned off by default after delivery, and only key login is allowed to reduce the risk of brute force cracking. If it has not been uploaded yet, you can submit the key in the "Instance Details" of the console and trigger the reinstallation process.
What system images are supported? How to reinstall?
The console provides a list of supported macOS versions (updated with Apple stable releases). Reinstalling will erasesystem diskPlease back up your data in advance. A work order can be issued for batch initialization of custom images and enterprises, and engineers will assist in verifying compatibility.
Hardware and underlying architecture
Are physical nodes 100% exclusive resources?
Yes. JEXCLOUD rejects any form of virtualization (No Hypervisor). Each delivery node isComplete Apple Silicon physical hardware. You have exclusive access to unified memory bandwidth, Neural Engine (NPU) computing power, and high-speed NVMe I/O without the interference of “noisy neighbors.”
How to choose M4 / M4 Pro / M4 Max?
combineUnified memory capacitySelect based on peripheral bandwidth requirements:Compilation and CIPrioritize larger memory and faster SSD;MLX inferencePrioritize larger memory to accommodate model weights. Specific specifications are based onPricing pageReal-time display shall prevail.
What is Thunderbolt 5 parallel technology?
For M4 Pro and above nodes, optional physical-grade Thunderbolt 5 (80 Gbps) links can be connected in parallel to connect multiple Macs toLow latency memory interconnectMethod synergy, suitable for large-scale LLM inference clusters or high-concurrency iOS compilation matrices.
Access and remote management
How to get SSH root privileges to log in?
Upload when placing an order or when the console resets the systemSSH Public Key. Use after delivery
ssh root@<Instance IP>Log in. Password login is disabled by default and only key authentication is supported.
Why doesn't VNC screen sharing connect?
For security reasons, the hardware firewall blocks public network inbound port 5900 by default. Please log in to the JEXCLOUD console and invoke the VNC desktop through the end-to-end encrypted Web tunnel; or add your public IP to the whitelist in "Network Security".
How does the console's "power management" work?
JEXCLOUD integrates enterprise-level MDM and computer roomPDU. Restart/shutdown commands on the web page can be issued to the underlying power module; even if macOS is unresponsive, a hardware-level restart can still be attempted.
Development environment and AI inference
How to configure the MLX large model inference environment?
Apple Silicon unified memory is a perfect fit for the MLX. suggestion:
- Use Conda or venv to manage Python environments;
- Install
mlx,mlx-lmWait for the official stack; - The physical machine can directly observe power consumption and heat dissipation, which facilitates long-term reasoning and optimization.
Is it possible to self-host a Runner as a GitHub Actions?
Can. Compared with cloud-shared Runners, exclusive M-series hosts usually have higher CPU/GPU utilization and stable disk IO, and are suitable as team-specific macOS CI nodes; just follow the GitHub documentation to register a Runner.
Any suggestions for Docker and Xcode build environments?
Docker Desktop for Mac can be used directly; for large iOS projects, it is recommended to reserve enough unified memory and enable incremental compilation. If the build time is concentrated in the linking phase, you can consider higher internal archive bits or split module builds.
Billing, compliance and support
What billing cycles are supported? How to renew?
supplyDaily rent, weekly rent, monthly rentWaiting period (subject to console order). Longer periods usually enjoy higher discounts. Please renew on the console before expiration to avoid triggering due to arrears.Resource recovery strategy(See detailsTerms of Service).
How do business users obtain invoices or sign contracts?
Enterprise customers can submit invoicing information and public payment vouchers on the console; large-amount or customized SLAs can sign offline contracts through the business/work order path. The specific process is subject to the agreement at the order stage.
How is the SLA guaranteed? What should I do if a long-term failure occurs?
we promise99.9%Level computer room power and network availability goals. If our infrastructure is unavailable for a long time, service time compensation or bill reduction will be provided according to the terms. Please pass the work order retention time and log voucher.
How do I contact support if I have a bill objection or technical glitch?
Please give priority to passconsoleSubmit a work order and attach the instance ID, time and screenshot to facilitate quick location by frontline colleagues. For emergency availability events, please escalate according to the emergency contact information announced on the console.
Where can I find the data and privacy policy?
Seeprivacy policy: It explains the log type, retention period, and how you manage the data yourself within the instance.Boundaries of responsibility.