By Erez Schnaider, Technical Product Marketing Manager, ClearML
The number of AI applications are rapidly increasing, and it can be difficult to keep up. Every month brings a new protocol, LLM, or tool. In this environment, the true strength of a platform is measured not only by its core features but also by its extensibility and adaptability to change. Many platforms address this challenge by hosting OSS tools or exposing API connections. ClearML goes further, offering a native plugin ecosystem that enables both ClearML and its customers to extend the platform’s capabilities with purpose-built applications. The result? ClearML customers skip months of custom development by deploying ClearML’s pre-built AI Applications.
Extending ClearML’s Platform with Applications
ClearML ships with a wide range of ready-to-use applications designed to expedite frequent tasks out of the box. These applications simplify workflows, improve visibility, and reduce manual effort in day-to-day AI project management. Some notable examples of applications include:
- Container Launcher – Automates the launch of containerized workloads, giving teams an easy way to spin up and manage environments on provisioned resources.
- Hyperparameter Optimization – Allows AI builders to perform optimal parameter search from within the ClearML UI.
Trigger Manager – Lets users configure event-driven automations, such as kicking off pipelines or sending alerts when specific conditions are met. - Quadrant / Milvus DB Deployment – Simplifies vector database setup, allowing users to simply deploy vector databases for advanced search and retrieval capabilities into their AI pipelines, especially for RAG applications.
For a full list of applications, check out this page within our documentation.
These applications are only a partial list of applications shipped natively within the ClearML platform and demonstrate the flexibility of ClearML: a single platform supporting both high-level project oversight and low-level workflow automation.
ClearML Applications are first-class citizens in the platform and inherit its core capabilities, security, and policies. They can run on ClearML-managed resources, respect multi-tenancy boundaries, and enforce strict role-based access control (RBAC). For multi-tenant environments, admins can monitor and attribute per-tenant billing events for both application usage and underlying compute consumption on top of ClearML-managed resources. Application metrics flow into ClearML Reports and can be linked to other tasks or entities, such as datasets or pipelines, across the platform.
Anatomy of an Application
Every ClearML Application follows a simple but powerful structure that makes it both extensible and approachable:
- Configuration Wizard – Provides an interface for user inputs, guiding users through configuration in a structured, accessible way.
- Dashboard – Displays the application’s outputs, whether metrics, logs, visualizations, or summaries.
- Python Code – Defines the application’s core logic and execution. This is where the actual workflow, logic, and visualization code lives. This can be an existing Python code modified to read from the Wizard and generate widgets to be displayed in the application’s dashboard.
This architecture ensures that applications can be as simple or as complex as needed (since they are written in Python), while still offering a consistent user experience that makes these custom workflows accessible to a wide range of users.
Custom Applications
ClearML doesn’t stop at providing pre-built applications. The platform also gives customers the ability to build a self-service catalog for custom capabilities, tailored to their unique workflows. Custom applications can further integrate ClearML into existing workflows and systems, including:
- Automating specific business or research processes and providing a useful user-interface for visualization results.
- Making technical workflows accessible to non-developers through guided interfaces.
- Providing custom dashboards and GUIs on top of ClearML’s existing infrastructure, and much more.
ClearML offers its customers documentation, including an application-building guide with example source code, to help teams build, deploy, and maintain these custom applications. By doing so, organizations can extend ClearML into a use-case specific toolset that matches their exact needs, while still benefiting from the scalability and governance of the core platform.
Wrapping Up
Applications are where ClearML’s flexibility really shines. The native plugin ecosystem allows teams to move beyond “using ClearML” into shaping ClearML into the platform they need. Whether through built-in apps like the Container Launcher and Trigger Manager, or through custom-built workflows and dashboards, ClearML Applications ensure that the platform evolves as quickly as AI itself.
Want to see it in action? Request a demo at clear.ml/demo