Rust-backed Excel engine for Python
For supported modify-mode workflows, WolfXL changes the targeted workbook data while preserving existing template parts that openpyxl's full-DOM save path can drop or rewrite, including slicers, some charts, and pivot layouts. Verified supported test cases open clean in Excel and LibreOffice, with supported-scope modify operations up to 45x faster, bulk reads up to 12x faster, and bulk writes up to 16x faster than openpyxl 3.1.5 (release-artifact benchmark, 2026-07-07).
Benchmarks: WolfXL vs openpyxl 3.1.5, clean-source release-artifact benchmark rerun dated 2026-07-07. See the compatibility matrix and limitations for exact scope.
-import openpyxl+import wolfxl as openpyxl# existing supported-surface code runs unchangedwb = openpyxl.load_workbook("report.xlsx")WolfXL implements the supported subset of the openpyxl API surface (load_workbook, Workbook, Font, PatternFill, and cell access patterns), so existing code on that surface runs unchanged. Update one import line or invoke the runtime alias at startup:
import wolfxl
wolfxl.install_as_openpyxl()The Rust core bypasses Python serialization overhead. Versus openpyxl 3.1.5 (clean-source release-artifact benchmark, 2026-07-07):
ws.append() and ws.write_rows().ws.iter_rows(values_only=True).Instead of rebuilding the whole workbook on save, modify mode parses the workbook, updates only the changed cells, and writes the ZIP parts back.
Construct interactive pivot tables for supported pivot-cache workflows. Unlike libraries that output flat, static cells, WolfXL writes a pre-populated pivotCacheRecords snapshot: verified test cases open in Excel or LibreOffice with pivot data already populated, no manual refresh needed.
Clean-source release-artifact benchmark dated 2026-07-07, openpyxl 3.1.5 vs WolfXL, supported scope.
Bulk reads
up to 12x faster
Bulk writes
up to 16x faster
Modify + save (2 cells, large workbook)
0.18s vs 9.7s
After checkout, your Cloudsmith private index token is provisioned manually and emailed to you. Review the terms, refund policy, license, and privacy policy before buying.
Licensed developers receive a private, tokenized index URL. To install, point pip at your private index using the --index-url flag:
pip install wolfxl --index-url https://dl.cloudsmith.io/<customer-token>/synthgl/wolfxl/python/simple/Using --index-url ensures all requests route through the private repository. Using --extra-index-url allows pip to search both public PyPI and the private index, creating a dependency-confusion risk where a public package could hijack the installation.
The WolfXL private package index proxies public PyPI. When you install WolfXL, the index server fetches and serves the public Python dependencies automatically.
Yes. Request a 30-day trial token and you get full access to the private package index: same wheels, same API, no feature gates. Email trial@wolfxl.com and we will review trial access during early access.
Early pre-launch development builds (through 2.0.0) were briefly published to PyPI under MIT before the commercial launch. Those releases are yanked, unsupported, and receive no fixes or updates. All supported versions are distributed via the private package index under the commercial license.
No. We do not include telemetry, phone-home license activation, or runtime usage tracking. We rely on standard organizational compliance. Your private index token regulates download access.
No. In licensed WolfXL 2.1.0+ distributions, confirmed correctness bugs in the supported openpyxl-compatible API surface are prioritized the same way for every licensed tier. Higher tiers buy faster support response targets, account help, deployment terms, and feature access. They do not buy access to bug fixes that lower tiers are denied. We do not withhold a confirmed correctness fix to push a customer into a higher plan.