Nobody uses Excel for just one thing. On one machine it is a personal calculator; shared across a team it quietly becomes a database with no types and no validation; wired into a business it turns into a reporting engine that emails XLSX files around. That range is why there is no single open source replacement, since the tools split by which job you are actually moving. Some are full office-suite spreadsheets that read your existing files, some add real columns and referential structure so a stray paste cannot corrupt a record, and some generate finished documents from data. Open source here means the format is documented, the data sits on hardware you control, and no license tier gates the features you rely on.
Grist is a modern relational spreadsheet for organizing data in tables that behave like spreadsheets but keep database-like columns and references. It is meant for people who need spreadsheet workflows with more structure for records, formulas, and shared data.
ONLYOFFICE Docs is a free collaborative online office suite with viewers and editors for text documents, spreadsheets, presentations, forms, PDFs, and diagrams. It is fully compatible with Office Open XML formats - .docx, .xlsx, and .pptx - and supports real-time collaborative editing.
LibreOffice is an integrated office suite for personal productivity. It is compatible with most document formats and standards, and includes applications for writing documents, working with spreadsheets, creating drawings, and building presentations.
EtherCalc is a web spreadsheet for real-time collaborative editing. It is a TypeScript rewrite of a Node.js port of Multi-user SocialCalc, built to let people share a room URL and edit the same sheet together.
Carbone turns JSON data into finished documents. You design a template in LibreOffice, Google Docs, or Microsoft Office and drop in JSON-like markers such as {d.companyName}; Carbone fills the markers from your data and outputs PDF, DOCX, XLSX, ODT, PPTX, ODS, XML, or CSV. The data comes straight from your existing JSON APIs.
Nurse Scheduling System solves the nurse scheduling problem, where constraints vary widely between hospitals and wards and manual scheduling in spreadsheets can take many hours. It aims to model real-world scenarios in one flexible framework rather than an over-simplified constraint set.
Pick by what your workbook is really doing.
Best when the spreadsheet is really a database: Grist Grist keeps the spreadsheet feel but backs it with database structure: typed columns, references between tables, and validation a plain grid cannot enforce. It supports Python formulas alongside Excel functions, has a REST API and webhooks, and stores each document as a portable SQLite-based file. Reach for it when a workbook has quietly become the team's system of record.
Best drop-in for everyday workbooks: LibreOffice LibreOffice Calc is the closest thing to a like-for-like desktop replacement, part of a full offline office suite that reads and writes most Excel files. It runs on Windows, macOS, and Linux with no account or subscription. Its StarBasic macros and UNO APIs give you a scripting path when you need to rebuild automation by hand.
Best for real-time collaborative editing: ONLYOFFICE Docs ONLYOFFICE Docs is built around real-time collaborative editing with strong Office Open XML compatibility, so .xlsx files round-trip more faithfully than most. It is self-hostable and covers documents, presentations, forms, and PDFs alongside spreadsheets, which suits teams that want several people editing the same sheet on infrastructure they control.
Best for generating documents from data: Carbone Carbone is not a spreadsheet at all; it turns JSON data into finished documents. You design a template in a spreadsheet or word processor and it outputs PDF, XLSX, DOCX, or CSV. It is the right tool when your spreadsheet is really a repeatable report or invoice generated from other systems.
The trouble with Excel is rarely the format, since XLSX is widely readable; it is what a spreadsheet becomes over time. A file that began as a quick calculation ends up as the team's database, edited by ten people, copied into five conflicting versions, with a formula silently wrong in one of them. So the first thing to settle is what your Excel is actually doing. If it is a personal calculator or a formatting-heavy report, a full office-suite spreadsheet like LibreOffice Calc or ONLYOFFICE that reads XLSX directly is the natural landing spot. If it has drifted into a shared system of record, a relational spreadsheet like Grist fits better, because it adds typed columns, validation, and references that stop a stray paste from corrupting data.
The gaps to test before committing are automation and the Microsoft-specific edges. VBA macros do not run unchanged anywhere here, so business-critical automation is effectively a rewrite: Grist offers Python formulas and LibreOffice has its own macro system, but neither is a drop-in for the VBA object model. Advanced pivot behavior, conditional formatting quirks, and locale-driven date handling can also differ enough to change a result, so compare outputs against Excel on representative inputs rather than trusting that a file which opened without an error also calculates the same.
For anything that is really a document job, such as invoices, contracts, or statements built from data, Carbone is a different shape of answer. You design the template in a spreadsheet or word processor, drop in markers, and it fills the values and exports PDF, XLSX, or DOCX, which is cleaner than maintaining a macro-driven master workbook.
Usually the common cases do. Simple workbooks with ordinary formulas, formatting, and charts open well in LibreOffice Calc or ONLYOFFICE, both of which read XLSX directly. The files that need real testing are the ones with macros, complex pivots, external data connections, protected ranges, or unusual conditional formatting. Open a copy, compare calculated results against Excel, and check anything you will need to send back to Excel users.
Assume they will not run unchanged. VBA's object model, event handling, user forms, and add-in dependencies do not carry over to another tool. LibreOffice has its own StarBasic macro system and Grist supports Python formulas, but both are a rewrite rather than a port. Treat any business-critical automation like a small software migration: document its inputs, outputs, and failure modes before rebuilding it.
Common formulas usually translate, but compatibility is not absolute. Differences show up in newer functions, array behavior, date systems, locale-specific separators, and rounding, any of which can change a result without throwing an error. For a workbook that drives financial or operational decisions, build test cases with known Excel outputs and compare, rather than trusting that a file which opened cleanly also calculates identically.
When the file has stopped being a calculation and become a shared record that people keep breaking. If several people edit it, a stray paste can corrupt a column, or you are tracking related lists that should link together, a relational tool like Grist fits better than any grid. It adds typed columns, validation, and references between tables while still letting you write formulas, including Python ones.
Yes, but it depends on the tool's model. ONLYOFFICE Docs is built around real-time collaborative editing with strong Office Open XML compatibility, so it comes closest to Excel coauthoring on infrastructure you host. Desktop suites like LibreOffice are stronger as single-editor applications with file-based sharing. If simultaneous editing is central to your workflow, test it with a realistic number of editors before committing.
You can, as long as you standardize on XLSX for anything shared with Excel users and test the round trip. LibreOffice and ONLYOFFICE both read and write Office formats, but heavily formatted workbooks, macros, and intricate pivots can shift on the way back. Save your own working copies in an open format for long-term control, and reserve XLSX for the files that actually cross the boundary.
Not as a direct feature match. You can import, clean, and reshape data in these tools, but the queries, refresh steps, relationships, and calculated measures usually have to be recreated rather than imported. If a workbook is really an analytics pipeline, consider whether a database plus a reporting tool fits better than another spreadsheet, and whether something like Carbone should handle the finished report output.
