GenOffice is an open-source AI-native office suite for macOS and Windows from the Genspark team. The repository describes five Electron apps sharing one engine layer: Docs for .docx, Sheets for .xlsx, Slides for .pptx, PDF for PDF editing, and a shell that hosts the suite. The important distinction is that AI editing is part of the document workflow rather than a side chat bolted onto a conventional editor.
The suite is designed around file fidelity. GenOffice Docs parses Word documents into block trees, tracks dirty paragraphs, regenerates only the changed OOXML fragments, and preserves untouched bytes from the original archive. Sheets builds on the open-source Univer core with in-house extensions plus a Rust sidecar for spreadsheet import and export. Slides uses an in-house PowerPoint parsing and rendering engine, while the PDF app combines pdf.js and pdf-lib for viewing, annotations, forms, signatures, and page operations.
Every app embeds a shared AI panel. In documents, that means block-granular editing with snapshots and diffs. In spreadsheets, slides, and PDFs, the README describes tool-calling agents over workbook, slide, and PDF state. Model calls route through the Genspark service side after sign-in, so users do not store their own model API key locally. For builders, the interesting part is the architecture: TypeScript engine packages, app-specific editors, shared provider abstractions, and testable parsing layers separated from Electron.
GenOffice is a strong fit for users who want AI assistance inside familiar office file formats without giving up round-trip compatibility. It is also useful for developers studying how to build AI-aware document editors, because the repository includes engine packages for DOCX, PPTX, file parsing, agent core, AI provider integration, search tools, i18n, UI, project storage, and Electron utilities.
Signed installers are published for Apple Silicon macOS and Windows x64, and the repository is licensed under Apache-2.0 with an enterprise-license exception for the reserved ee directory. Pricing for the open-source desktop suite is free at the code level; users should still expect Genspark account requirements for AI provider routing and any future commercial modules that the maintainers decide to ship.
The strongest reason to watch GenOffice is not only that it is an office suite with AI. It is that the repository exposes the engineering seams. The DOCX engine tracks original XML slices and only patches dirty blocks. The spreadsheet app combines an open-source core with an in-house Rust sidecar for import and export. The presentation engine handles PowerPoint structure directly, and the PDF app works with known open libraries. That gives developers a concrete example of how AI features can be inserted into document workflows without treating the file as disposable text. For teams that exchange files with Microsoft Office users, byte preservation and narrow patching matter. For open-source contributors, the separated packages make it possible to inspect one layer at a time.