What’s New in PyCharm 2025.2

In our efforts to make PyCharm one powerful unified product, 2025.2 will be the last supported version of PyCharm Community. With the 2025.3 release, we’ll offer a seamless way to migrate to the unified PyCharm. You can upgrade to 2025.2 Community using the Toolbox App or via the in-product upgrade notification.

If you’d already like to start using the unified PyCharm with Jupyter notebooks for free today, without waiting for the migration, you can download it here.

Learn more about the unified PyCharm

PyCharm 2025.2 introduces the versatile new AI Playground feature, which lets AI engineers compare responses from multiple AI models side by side. You can configure system prompts and fine-tune individual model parameters, enabling precise control over Temperature, Top P, and Max length parameters.

Go to More Tool Windows | AI Playground, where you will be able to choose between:

• All models included in your paid AI subscription (these will use your AI Assistant tokens^Experimental).
• Custom providers that you can add via API keys.
• Locally running models.

PyCharm 2025.2 also introduces the powerful new AI Agents Debugger plugin, designed to provide transparency and debugging capabilities for AI agent development. It’s built for developers working with agentic systems who need to understand, debug, and optimize their agent workflows.

AI Agents Debugger transforms the typically opaque "black box" nature of agentic systems by providing detailed insights into the thought processes, metadata, inputs, and outputs of each agent node.

Junie, the coding agent by JetBrains, autonomously plans, writes, refines, and tests code to make your development experience smooth, efficient, and enjoyable. It handles tedious tasks like restructuring code, creating tests, and implementing refinements, so you can focus on bigger challenges and innovation.

Junie has received several major updates, including a speed boost of up to 30%! It now features MCP support, enabling Junie to connect to external servers and deliver more accurate solutions. Additionally, Junie is now available for remote development (currently supported on macOS and Linux) and has launched an Early Access Program for its GitHub integration. This integration allows you to work with pull requests without opening the IDE.

Junie is available in PyCharm as part of the JetBrains AI subscription.

• Smarter completion across all supported languages, now with support for SQL, YAML, JSON, Markdown, and more.
• Project rules mean you can instruct AI to comply with your team’s coding conventions and business logic.
• Expanded offline flexibility – connect any OpenAI-compatible model server like llama.cpp or LiteLLM.
• JetBrains AI Assistant now supports image attachments in the chat when using Anthropic and OpenAI models.

All JetBrains AI features are available for free with a PyCharm Pro subscription, with unlimited code completion, powerful local workflows, and limited cloud-based features.

PyCharm now supports one-click setup for the MCP (Model Context Protocol) server, making it easy to connect external AI clients like Claude, Cursor, or Windsurf. When a compatible client is detected on your machine, your IDE will prompt you to enable the server and auto-configure the connection – no manual setup required.

With this integration, your preferred AI clients' capabilities are extended with the intelligence of PyCharm, letting you trigger actions like running unit tests, refactoring code, and generating files.

We’ve added persistent controls to .toml configuration files for uv, Hatch, and Poetry. These controls allow you to apply changes made to a .toml file by clicking the corresponding buttons to sync, lock, or update the dependencies in your environment.

You can now generate code for Jupyter notebooks and SQL cells using AI Assistant, which considers the source cell’s context.

To do so, click AI Edit Cell in the context menu or navigate to it from the AI icon in the right-hand toolbar.

Logical structure offers a framework-centric view of your Django project, allowing you to navigate and understand your application architecture more intuitively. It represents your project based on how Django sees it – surfacing the meaningful relationships between models, views, serializers, forms, admin classes, and other components you work with every day. It also allows you to generate templates for missing relations between entities, helping you quickly scaffold your code.

With an active JetBrains AI subscription, you can unlock more advanced capabilities. Instead of basic templates, the AI generates complete code that’s fully aware of entity relationships, delivering smarter and more context-rich results.

A new image viewer provides options for performing additional manipulations, including restoring the original image, inverting the channels of the original image, converting the image to grayscale, setting the binarization threshold, and more.

We've introduced several exciting updates to enhance your database experience! Editable results for SELECT queries with JOIN clauses are now supported, allowing you to directly modify query results in the grid. PostgreSQL and Microsoft SQL Server benefit from introspection by levels for faster metadata loading, while PostgreSQL also gains a smart refresh mechanism to update only affected objects after DDL changes. Also, SQLite now supports WSL file paths, resolving write-locking issues, and MongoDB connections now allow you to configure read preferences and replica set connectivity.

Read about these features in more detail.