VS Code C/C++ IntelliSense Cache
The Microsoft C/C++ extension for VS Code stores IntelliSense precompiled header data in its ipch cache so it can parse large codebases faster without reprocessing the same headers on every file open. Those ipch files contain precompiled translation-unit state derived from include paths, macros, compiler settings, and header contents, and they can become oversized or stale after SDK, toolchain, or configuration changes. Kudu removes only the VS Code C/C++ IntelliSense ipch cache so the extension can rebuild it automatically the next time source files are opened, without touching source code, settings, accounts, or workspace data.
Why clean VS Code C/C++ IntelliSense Cache?
- Changed include paths, defines, or compiler mode leave old ipch data out of sync, which shows up as false red squiggles and incorrect unresolved symbol errors until IntelliSense reparses
- After a Windows SDK, Xcode Command Line Tools, or GCC/Clang update, cached precompiled state can reference older headers and cause autocomplete to suggest the wrong APIs or miss new ones
- A bloated ipch folder from large workspaces consumes several gigabytes under the user cache directory, leaving users wondering why VS Code keeps growing on disk
- Corrupted precompiled header chunks can make file opens stall while the C/C++ extension repeatedly retries parsing, which feels like IntelliSense is stuck on 'Updating' or never becomes ready
- Switching branches or generated headers without a clean rebuild can leave stale macro expansions in cache, causing Go to Definition and hover type info to disagree with the actual build
- On slower disks, fragmented and excessive ipch files increase random reads during parsing, so code completion appears late and navigation features lag when opening common headers
Cache paths Kudu targets
Windows
%LocalAppData%/Microsoft/vscode-cpptools/ipch |
macOS
~/Library/Caches/vscode-cpptools/ipch |
Linux
~/.cache/vscode-cpptools/ipch |
Common questions about VS Code C/C++ IntelliSense Cache
Download Kudu and reclaim your disk space.
Available on Windows, macOS, and Linux. No account required, no feature gates, no telemetry without consent. All cleaning targets are open source and community-auditable.