CloudPilot Feature Overview: Managing AWS, Azure, Cloud Browsers, and AI Workspaces in One Web Page
CloudPilot is a cloud resource management panel I developed that brings AWS, Azure, cloud browsers, remote desktops, and AI workspaces into a single web page. It can manage servers across multiple accounts and regions, and it can also record and replay browser operations, reducing the back-and-forth switching involved in deployment, testing, development, and resource cleanup. This article uses feature screenshots to explain what it can do and what usage scenarios it suits.
If you use several cloud accounts at the same time, you may have run into situations like these: creating a machine requires switching consoles, connecting to a desktop requires opening another piece of software, and running automation requires setting up an environment separately. CloudPilot's starting point is to bring these frequently co-used operations together.
Project entry point:Open CloudPilot; usage instructions:Features and Tutorials。
The images in this article come from the demo environment of the program's actual interface. Accounts, IPs, resource names, costs, and workspace data are all examples; they do not represent real accounts, nor do they represent default quotas or actual prices. The interface and features will continue to be updated with new versions. Click on images to enlarge them.
1. Multi-account management: switch credentials without repeatedly entering keys
CloudPilot supports saving multiple sets of AWS credentials and provides search, sorting, batch import, and status checks. When there are many accounts, you can name them by project, such as development, testing, and production, and then switch the currently used credentials on the corresponding feature pages.
Azure, meanwhile, manages access information by account and subscription. The panel also provides a CloudShell onboarding guide: running the command generated by the panel in an account that supports CloudShell can simplify the process of creating and adding credentials. After onboarding, resource operations are still limited by the permissions of the corresponding cloud account.

2. AWS: from creating instances to remote management
The AWS section covers EC2, Lightsail, and Wavelength-related operations. When creating instances, you can select the region, operating system, specifications, and quantity, and you can also configure deployment targets across credentials and regions, reducing the work of repeatedly filling out forms for each account.
The instance list puts running status, region, specifications, and network addresses together. After selecting multiple machines, you can batch start, stop, restart, manage ports, or open web-based RDP. For those who frequently create test environments, centralized viewing and batch operations are more convenient.
Windows Remote Desktop can be opened directly from the web page. When the same operations need to be performed in multiple Windows environments, you can also use synchronized operations, desktop recording, and background replay. It suits scenarios such as repeated configuration, software demonstrations, and test preparation.

3. Azure: virtual machine creation and subsequent maintenance
Azure virtual machine management has been integrated into the panel, including account onboarding, region and specification selection, Linux and Windows images, batch creation, start/stop/restart, connection information, and task records.
The creation window pre-fills common fields such as name, operating system, and username, and retrieves available regions and VM specifications based on the subscription. Each deployment target can have its account, region, and quantity set separately, and vCPU and memory are also displayed next to specifications for easy direct comparison.


For student subscriptions, the panel provides filtering and recommendations for free-tier eligible machine types, and cloud browser deployment has also been adapted for low-memory machines. However, "belongs to a free-tier eligible machine type" does not mean all resources are free of charge: whether credit is available, usage duration, and costs such as disks and public IPs should still be based on subscription benefits and the Azure bill.
When deleting virtual machines created by the panel, you can choose to clean up the machine and associated resources. For temporary test environments, this can reduce situations where only the virtual machine is deleted while disks or network resources are left behind.
4. Cloud browsers: use independent browser environments within a web page
Cloud browsers can be used for website testing, remote browsing, and work environments that need to retain configuration over the long term. CloudPilot offers browser options such as Brave and Chromium, and supports Camoufox in the corresponding environments; the specific available options depend on the deployment method.
You can use hosted cloud browsers, or deploy browsers to your own EC2 or Azure Linux virtual machines, or bind an existing Linux VPS. During creation, you can configure a URL to open automatically, additional bookmarks, and compatible browser extensions, reducing the steps of repeated setup each time you enter.
The browser list supports batch opening of desktops, synchronized operations, recording and replay, and unbinding. For self-hosted environments, unbinding from the panel and deleting the cloud server are two different operations, and you need to choose according to the actual purpose.

5. Recording, scheduled replay, and Playwright scripts
The automation section offers two approaches: one records the operation process on the desktop and then replays it on a schedule; the other uses Playwright to perform clicks, inputs, waits, and assertions through page element locators.
The Playwright panel supports using the official recorder to generate code, and can also import or edit JavaScript and TypeScript test scripts. Scripts can be saved to an account, and you can select multiple execution targets and set the start delay, number of runs, interval between runs, and per-run timeout.

For example, when testing your own website, you can first record the flow of "open the homepage, enter the form, fill in test content, check the result," and then hand the script to multiple browser environments for execution. When more complex judgments are needed, you can continue editing the code. The script runtime environment has module and system access restrictions and is not equivalent to a server that can run arbitrary Node.js programs.
The loop workflow, meanwhile, chains together environment creation, waiting, replay, and cleanup for the current round, making it suitable for tests that require repeatedly setting up temporary environments. Desktop replay relies more on interface layout, while Playwright relies more on page locators; in actual use, you can choose according to the task.
6. Cloud Agent: put programming environments and desktops in the cloud
In addition to cloud resource management, CloudPilot also provides cloud workspaces. The AI programming workspace comes with Codex and Claude Code built in, and you can use the corresponding service's account or API Key; the desktop workspace provides Ubuntu, Chrome, and DeepSeek Harness, and can be accessed from the web page.
Workspaces can be created, named, and opened separately by project. They sleep when idle and wake up when you need to continue working, making it suitable for continuing work on the same project across different devices.

Workspaces provide entry points such as files, Git, terminal, and settings. The file manager supports upload, download, and creating new files and folders, and you can also enter a path directly to locate something—for example, paste the path of a ZIP generated by the project into the address bar, then jump to the file's location.

Automatic backups during operation, restore points, and records of exit reasons help with recovery and troubleshooting after abnormal stops. For important projects, it is still recommended to also commit to your own Git repository or download archives regularly, so as not to keep all data in only one workspace.
7. Billing, Credit, and scheduled tasks
When managing cloud servers, costs also need to be kept in view. The AWS billing page can show cost summaries and service details; when you have the corresponding permissions, it can also read Credit balances, expiration dates, and applicable services, and provides a Credit Code redemption entry point.
This data comes from AWS and may have update delays, and may also be affected by IAM, billing access settings, and organizational policies. The panel displays the returned results together for easy viewing across different accounts.

Scheduled tasks support scheduling instance operations once, daily, weekly, or monthly, and display the time zone and execution results. For example, you can set a schedule for a weekday test machine to start in the morning and stop in the evening, reducing situations where you forget to deal with the machine.

8. Resource cleanup, CloudFront, and common maintenance tools
Resource cleanup first displays the results to be processed, including unassociated EC2 elastic IPs, Lightsail static IPs, and available volumes and unreferenced snapshots within the panel's management scope, so you can confirm before cleaning up.

The CloudFront page centrally provides distribution creation, origin and custom domain configuration, ACM certificate guidance, distribution status, and cache invalidation. For users running websites, static assets, or test origins, this reduces jumping between different console pages.

Other common tools also include AMI management, region status management, vCPU quota queries, and Wavelength environment configuration. These features revolve around the same purpose: after an environment is created, subsequent maintenance can also be completed in the panel.
How to get started
- Open CloudPilot, and first look at the demo interface and feature descriptions.
- Register and log in, and add AWS credentials or an Azure subscription as needed.
- Choose a practical task to start with, such as managing existing instances, creating a test machine, or opening a cloud workspace.
- In Features and Tutorials , check the corresponding operation steps, then gradually use batch deployment and automation.
CloudPilot is an independently developed management panel and is not an official AWS or Microsoft console. The terms of use for the panel service are calculated separately from cloud provider resource costs; cloud resources created through the panel are still billed in your own account. For actual features, quotas, and available regions, please refer to the information returned by the current panel and cloud account.
If you are managing AWS and Azure at the same time, or want to bring remote browsers, automated testing, and AI programming workspaces together into one entry point, you can start here:Try CloudPilot。
评论与讨论
还没有评论,欢迎留下第一条讨论