AirCloud April Update
- Apr 29
- 4 min read
AirCloud's April release is built around one goal: making it faster to run AI workloads, more reliable to operate them, and more flexible to put your existing GPU resources to work.
This update includes enhanced Air Container operations, the general availability of Air API, Resource Provider (RP) support, and the introduction of an intelligent scheduler. Developers can now handle container access, log monitoring, error response, and API integration more seamlessly. Enterprises, research institutions, and GPU owners can connect their existing hardware directly to the AirCloud resource pool.
1. Air Container: Stronger Operations
When running AI workloads, the tasks that come up most often are connecting to a container, checking logs, and diagnosing issues quickly. This release improves Air Container across three dimensions: accessibility, operational visibility, and control.
SSH Access
Air Container now supports two ways to connect. Web-based SSH lets you access your container directly from the browser with no additional setup. Direct SSH terminal access connects from your existing local development environment. Choose whichever fits the situation and get to debugging faster.
Improved Log History Navigation
Finding the right log entry from a past run is now easier. The container execution history and log list have been redesigned for better visibility, so when something goes wrong, you can locate the information you need without hunting through unstructured output.
Enhanced Error Notifications
Delayed awareness of an error means delayed response. The error notification flow has been improved so that issues surface more clearly during workload execution, giving you a faster starting point for any follow-up action.
AirCloud Control CLI
AirCloud resources can now be managed from the CLI, not just the console. This makes it easier to automate repetitive tasks, build operational scripts, and integrate AirCloud into existing developer workflows.
Shared Storage
When multiple containers or workloads need to access the same data, repeatedly uploading and downloading model files, datasets, or output artifacts creates unnecessary overhead. Shared Storage lets you keep data in one place and access it across workloads without duplication.
2. Air API: Now Generally Available
Air API is now officially GA.
Air API provides LLM and TTS capabilities through an OpenAI-compatible interface. Developers already familiar with OpenAI-style integrations can switch to Air API without reworking their existing setup. The API is designed to balance performance, latency, cost efficiency, and reliability.
Use cases include chatbots, AI agents, document summarization, code generation, data analysis assistance, and voice content generation. Generate an API key from the AirCloud console and start integrating immediately. LLM and TTS APIs are available within the same Air API environment.
💬 LLM API Models
Qwen: Qwen3.5-35B-A3B
Qwen: Qwen3.5-9B
🗣️ TTS API Models
Qwen: Qwen3-TTS
3. Resource Provider: Connect Your GPU Infrastructure to AirCloud
AirCloud Resource Provider (RP) allows enterprises, research institutions, and GPU owners to connect their existing hardware directly to the AirCloud resource pool.
With this update, on-premise GPU servers, R&D equipment, idle GPU hardware, and infrastructure owned by GPU-holding businesses can all be extended into AirCloud's workload execution environment. This creates a foundation for managing previously fragmented GPU resources in a more structured way and putting them to productive use.
Connecting Existing GPU Resources as an RP
Any organization with GPU infrastructure can register as a Resource Provider and connect their hardware to AirCloud. Connected resources become available for workload execution, including hardware that would otherwise sit idle during off-peak hours.
This is relevant for a wide range of organizations: gaming cafes, research institutes, small and medium-sized businesses, and public institutions. Existing infrastructure that was previously underutilized can be activated as a computing asset within AirCloud.
Unified Dashboard and Monitoring
A unified dashboard gives Resource Providers visibility into the status, usage, and operational metrics of their connected GPU resources. Operators can monitor connected hardware and understand how it is being used at a glance.
Usage-Based Compensation
When GPU resources contributed by a Resource Provider are used to run AirCloud workloads, compensation is provided based on actual usage. This allows GPU-holding organizations to increase the return on existing infrastructure through a usage-based model.
4. Intelligent Scheduler
AirCloud now includes an intelligent scheduler.
The new scheduling engine analyzes workload requirements and the available execution environment, then automatically assigns the most suitable device. Users no longer need to manually compare GPU types or configurations every time they run a workload.
Automatic Device Assignment
The scheduler evaluates what a workload needs and allocates an appropriate device automatically. Getting the right execution environment requires no manual comparison on the user's side.
Reduced Cold Start Times
One of the more friction-heavy parts of repeated workload runs or service restarts is waiting for container images to download. The intelligent scheduler prioritizes devices that already have the relevant image layer cache, which reduces download time and shortens startup duration for recurring workloads.
Wrapping Up
The AirCloud April update is focused on moving AI workloads from slower and more manual to faster and more automated.
Developers get a better container experience and a production-ready API. Enterprises and GPU owners get a way to connect existing infrastructure to a broader resource pool. And the intelligent scheduler ensures that workloads land on the right device from the start, with faster startup times for repeated runs.
Check out the latest features in the AirCloud console and start building with Air API today.
