New Features in Python VS Code Extension (March 2026): Q&A

The March 2026 release of the Python extension for Visual Studio Code brings two notable enhancements to developer productivity. First, you can now search for symbols from installed packages directly within your workspace, making code navigation smoother. Second, an experimental Rust-based parallel indexer promises dramatically faster completions and symbol searches, especially on large projects. Below, we answer common questions about these updates, how to enable them, and what to expect.

What are the main new features in the March 2026 Python extension release?

This release introduces two key improvements to the Pylance language server within VS Code:

• Symbol search in installed packages – when enabled, Workspace Symbol Search (Cmd+T / Ctrl+T) includes functions, classes, and other symbols from Python packages in your active virtual environment. This helps you explore third-party libraries without leaving the editor.
• Experimental Rust-based parallel indexer – a new indexing engine that runs out-of-process and in parallel. In testing, it delivered up to 10× faster indexing on large projects, leading to quicker autocompletions and more responsive IntelliSense after workspace open.

Both features are opt‑in. You control their behavior through dedicated settings. For full details, see the changelogs for the Python and Pylance extensions.

How do I search for Python symbols across installed packages in VS Code?

To include symbols from your virtual environment’s site-packages in the Workspace Symbol Search, you need to enable the feature via a new setting. Go to File > Preferences > Settings (or press Cmd+, on macOS / Ctrl+, on Windows/Linux). Search for Include Venv In Workspace Symbols and check the box under Python > Analysis.

Once activated, pressing Cmd+T / Ctrl+T will show results from both your project code and the installed packages in your active environment. For libraries that do not include a py.typed marker, only symbols exported via __init__.py or __all__ are shown, keeping results focused. If you need finer control, adjust Package Index Depths (also under Python > Analysis) to limit how deeply Pylance searches into sub‑modules.

How much faster is the new Rust-based parallel indexer?

In our internal tests, the experimental parallel indexer is on average 10× faster than the previous indexing engine when working on large Python projects. This translates to noticeably quicker completions, auto‑imports, and workspace symbol searches right after opening a workspace. The improvement stems from the indexer running out‑of‑process and in parallel, leveraging Rust’s performance characteristics.

It is important to note that the speed gain is most pronounced on larger codebases. Small projects may see little to no difference because the overhead of spawning a separate process and indexing a small number of files is comparatively high. The feature is intentionally experimental: we want to validate performance and reliability across a wide variety of project setups before making it the default. Your feedback helps us decide when to promote it to stable.

How do I enable the experimental parallel indexer?

To try the new indexer, open VS Code Settings (Cmd+, / Ctrl+,), search for Parallel Indexing, and check Enable Parallel Indexing (Experimental) under Python > Analysis. Alternatively, add this line to your settings.json file:

"python.analysis.enableParallelIndexing": true

After enabling, you must reload VS Code (Cmd+Shift+P / Ctrl+Shift+P → Reload Window) to ensure the new indexer starts cleanly. The change takes effect immediately after reload. If you later want to disable it, uncheck the setting or remove the line from settings.json and reload again.

Will the parallel indexer work with all Python projects?

The new indexer is designed to work with any Python project that Pylance currently supports, but its benefits vary. As noted, large projects with many files and heavy package dependencies see the most dramatic improvement – often up to 10× faster indexing. For smaller projects, the performance difference may be negligible because the total indexing time is already short.

Because this is an experimental setting, we advise testing it on your specific project before relying on it in production. Some edge cases (e.g., unusual virtual environment configurations, projects with deeply nested symbolic links) might exhibit different behavior. We encourage you to report any issues or observations via the Pylance issue tracker. Your feedback will help us improve reliability and make an informed decision about promoting the feature to default status.

How can I provide feedback on these new features?

We welcome your input on both the symbol search enhancement and the parallel indexer. The best way to share your experience is through the official Pylance GitHub repository, where you can open a new issue or comment on existing ones. When reporting about the parallel indexer, please include details about your project size, operating system, and any notable changes in performance – faster completions, slower behavior, or crashes.

For general questions or discussions, you can also use the Python extension repository or the VS Code Python tag on Stack Overflow. By engaging with the community, you help us shape future releases and ensure that the most impactful improvements reach all users.