Release Notes¶

This page outlines key feature updates for d.run.

2025-06-29¶

[Container Instance] Support remounting file mount paths when changing configurations
[Compute Cloud] Integrated with audit logging
[Compute Cloud] Support retrieving operational metrics such as GPU utilization and memory usage for all GPUs under a container
[Container Registry] Added duplicate username check for container registry accounts

[Container Instance] Improved delay in status display after shutting down and saving the system disk
[Container Instance] Optimized disk limits for container instances
[Container Instance] Improved image save prompts for scheduled or manual shutdowns
[File Storage] Enhanced status display during file storage initialization

[Container Instance] Fixed issues when accessing Muxi container instances via JupyterLab, including terminal errors and missing GPU model info
[Container Instance] Fixed failure to start container instances created with custom image files
[File Storage] Fixed partial failures when uploading multiple files
[Container Registry] Fixed issue of two image versions appearing when manually shutting down and saving a system disk for the first time
[Container Registry] Fixed failure to save images when a container instance is shut down due to overdue payments

[Container Instances] Added monitoring support for domestic GPUs (Metax, Enflame, Biren).
[Container Instances] Enabled container instances to use custom images.
[Container Instances] Added a prompt for VRAM requirements of the image version when creating container instances.
[Container Registry] Introduced a container registry module to support image lifecycle management.
[Container Registry] Enabled manual saving of container instance images.
[Container Registry] Added support for changing the container registry password.
[File Storage] After 15 days of overdue payment, file storage will be automatically downsized to the free 20GB tier and users will be notified via SMS.
[Compute Cloud Admin View] Added support for linking the gateway protocol, networkSupported, and SERVICE_PROTOCOL parameters—only one of these needs to be configured.

[Container Instances] Improved delay in container instance status display after shutdown and system disk saving.
[File Storage] Reduced display delay after successful file storage scaling operations.

[Container Instances] Fixed an issue where data was lost when restarting a container instance after enabling scheduled shutdown and saving the system disk.
[Container Instances] Fixed display issues on the welcome page.
[File Storage] Fixed an issue where expansion orders were still generated after a failed expansion of Suoyuan (Enflame) file storage.

[Model Gallery] Added support for "Deep Thinking" in text models
[Model Gallery] Enabled message copying and regeneration for text models
[Model Gallery] Image-to-text models can now generate multiple images simultaneously
[Model Gallery] Image-to-text models now support custom positive/negative prompts and image sizes
[Served Models] Fully compatible with the standard OpenAI SDK
[API Keys] Usage statistics can now be filtered by API key, model type, and invocation time
[API Keys] Quickly view total calls, total input tokens, and total output tokens
[API Keys] Compare usage across different models

[Compute Market] Quickly purchase various types of GPU resources from the marketplace
[Container Instance] Lifecycle management (create/delete/start/stop) supported
[Container Instance] Optional Jupyter installation during creation, with quick access via Jupyter
[Container Instance] Optional VSCode installation during creation, with quick access via VSCode
[Container Instance] Support mounting file storage when creating instances
[Container Instance] Login and access via SSH with username and password
[Container Instance] Password-free SSH login via public key supported
[Container Instance] Pay-as-you-go billing available for container instances
[Container Instance] Create instances using built-in images
[Container Instance] Open instance ports via UI
[File Storage] Initialize file storage in different regions
[File Storage] Each region provides 20GB of free storage
[File Storage] Upload, download, and delete files via the UI

[Model Gallery] Card view for model list for better visualization
[Model Gallery] Detailed model descriptions and API call examples
[Model Gallery] Quick deployment and testing of text models
[Model Gallery] Search models by name, provider, and type
[Model Gallery] Try multimodal models
[Model Gallery] Try image-to-text models
[Served Models] View list of deployed model services
[Served Models] Variety of models available for deployment
[Served Models] Instance scaling support
[Served Models] Postpaid billing (pay-as-you-go) with cost transparency
[Served Models] Region selection supported for deployment
[Served Models] Configure number of instances for horizontal scaling
[Served Models] In-browser experience of deployed models
[Served Models] API examples in multiple languages (curl, Python, Node.js)
[Playground] One-click testing to quickly verify service availability
[Playground] Text generation models support parameters like system prompt, temperature, and top_p
[Playground] Compare different models of the same type
[API Keys] View API key list
[API Keys] One-click copy of full API key
[API Keys] Delete specific keys (irreversible)
[API Keys] Generate new API keys with custom names

[Wallet] Recharge with Alipay
[Wallet] View current cash balance
[Orders] View all account transactions (recharge, consumption, refunds) on a timeline
[Orders] View order details including status, time, and amount
[Bills] View charges and billing records for all created resources

Login/Registration – Sign up using mobile number and verification code
Login/Registration – Login with either username/email + password or phone + code
Password Recovery – Reset password using mobile number and verification code