douglas168/open-genoffice

GenOffice fork with a provider-agnostic AI slot — bring your own local LLM or OpenRouter key

stars

commits

TypeScript

primary language

Aug 18, 2026

updated

ai-agent

apache2

desktop-app

document-editor

docx

electron

llm

lm-studio

local-llm

microsoft-office-alternative

office-suite

ollama

openrouter

open-source

pptx

presentation-software

self-hosted

spreadsheet

word-processor

xlsx

README

GenOffice (local-LLM fork)

A free, open-source AI Office suite — this fork drops the cloud-account requirement and talks to any OpenAI-compatible endpoint instead: a local server (Ollama, LM Studio, vLLM, llama.cpp server, text-generation-webui) or a hosted API (OpenRouter, or any other OpenAI-compatible provider).

Forked from genspark-ai/genoffice (Apache-2.0). Changed per Apache-2.0 §4(b): the AI provider defaults (packages/ai-provider/src/providers.ts) and settings UI (apps/shell/src/renderer/src/SettingsModal.tsx) — see git history for the full diff. Everything else is upstream GenOffice, unmodified.

Demo — upstream's video; the editor features it shows are unchanged in this fork, only the AI provider setup differs.

GenOffice is a free, open-source alternative to Microsoft Office for macOS, Windows, and Linux, built around AI editing as a first-class workflow rather than a bolted-on chat box. It opens and saves the real Microsoft Office formats — Word (.docx), Excel (.xlsx), PowerPoint (.pptx) — and edits PDF and Markdown too: a word processor, spreadsheet, presentation editor, PDF editor, and Markdown editor as six Electron apps sharing one engine layer.

Watch the demo video on YouTube

Features

Real PDF editing — retype text and edit images in the page itself, original fonts preserved.
Microsoft Word–compatible, byte-preserving .docx editing — only what you touched changes; Word never notices.
Word-faithful pagination — page breaks land where Word puts them.
Excel-compatible spreadsheets — in-house engine with a Rust .xlsx sidecar, own charts, pivot tables, slicers.
PowerPoint-compatible presentations — in-house .pptx engine with masters, layouts, smart guides, non-destructive crop.
Markdown to Word, fully local — the same OOXML engine, no Pandoc, no cloud.
AI that edits documents — block-level edits with snapshots and diffs, document-aware agents.
Agent tools built in — web/image search, image generation, media analysis.
Light / dark / system themes.
macOS, Windows, Linux.
Free & open-source (Apache-2.0).

Get it

This fork does not publish prebuilt installers — build from source (see Development below: npm install, then npm run dist:mac / dist:win / dist:linux). Upstream's own prebuilt installers, linked from genspark-ai/genoffice, default to the Genspark cloud provider — they are not this fork.

Apps

App	Product	What it is
`apps/docs`	GenOffice Docs	`.docx` word processor. Byte-preserving round trip: only dirty paragraphs are regenerated (paragraph patch), everything else in the original file is kept byte-for-byte, so opening and saving never breaks layout in Word. Paginated view whose line metrics reproduce the original document's layout, tracked changes, comments, styles, equations, ink.
`apps/sheets`	GenOffice Sheets	`.xlsx` spreadsheet. UI built on the open-source Univer core (Apache-2.0) with a large layer of in-house extensions; `.xlsx` import/export runs through an in-house Rust sidecar (calamine + IronCalc), charts are rendered in-house (Konva), plus pivot tables, slicers, conditional formatting, and formula tracing.
`apps/slides`	GenOffice Slides	`.pptx` presentations. In-house `.pptx` parse/render/edit engine with masters, charts, cropping, ink, and text shaping (HarfBuzz metrics).
`apps/pdf`	GenOffice PDF	`.pdf` viewer/editor on pdf.js (Apache-2.0) + pdf-lib (MIT): annotations, forms, outlines, stamps, signatures, page operations, and printing support. True text editing — paragraph selection with in-block reflow, alignment restoration, original-font preservation — and content-stream image insert/edit, all rewriting page content streams through PDFium wasm (BSD-3-Clause) with subset-embedded fonts — no cover-up annotations.
`apps/markdown`	GenOffice Markdown	`.md` / `.markdown` editor: Tiptap block editor over plain Markdown files — headings, lists, tables, images, code blocks — saved back as plain Markdown, hosted in shell tabs.
`apps/shell`	GenOffice	The suite shell: home screen, tabbed hosting of the five editors, light/dark/system theme, auto-update.

Every app embeds the same AI panel: block-granular AI editing with version snapshots and diffs in docs, a tool-calling agent over workbook/slide/PDF state in the others.

The whole suite ships light / dark / system UI themes built on shared design tokens (packages/ui), with a CI guard that keeps chrome colors on the token system. Document surfaces stay light in dark mode — Word-style dark chrome around white paper — so files render and export identically in both themes.

AI backend. On a fresh install this fork defaults to the Local / Custom (OpenAI-compatible) provider with the base URL, model, and key all blank — nothing is pre-configured, no account or cloud sign-in required. Point it at any server that speaks the OpenAI chat/completions wire protocol:

Server	Base URL	Model	API key
Ollama	`http://localhost:11434/v1`	e.g. `llama3.1`	leave blank
LM Studio	`http://localhost:1234/v1`	the model loaded in-app	leave blank
OpenRouter	`https://openrouter.ai/api/v1`	e.g. `openai/gpt-4o-mini`	your OpenRouter key

Settings → AI Provider → Local / Custom (OpenAI-compatible), fill in the three fields, then Test connection.

Upstream's other providers are still selectable in the same screen and work as upstream describes: Genspark signs in via a device-code flow (no key needed) and also unlocks the Genspark ("gsk") tool endpoints the agents build on — web/image search, image generation and editing, media analysis, audio transcription (packages/ai-search); Claude / Gemini / DeepSeek / OpenAI need that vendor's own API key.

Engine packages

All pure TypeScript, no Electron dependency, unit-tested (except the UI kit):

packages/docx-engine — docx parsing → block tree (with docxIndex anchors and passthrough), OOXML fragment generation, byte-level paragraph patching.
packages/pptx-engine / packages/pptx-render — pptx model and rendering.
packages/file-parse — text extraction for AI attachments (office formats, text formats).
packages/agent-core — the AI agent loop and skill composition shared by every app.
packages/ai-provider — provider abstraction and streaming for the model backends.
packages/ai-search — Genspark auth + web/image search tools.
packages/i18n, packages/ui, packages/project-store, packages/electron-utils — shared i18n core, React UI kit, recent-files store, and Electron main-process helpers.

Development

npm install
npm run fixtures     # generate test .docx fixtures
npm test             # engine + app unit tests (docs/sheets/slides need no display)
npm run typecheck    # tsc --noEmit across every workspace
npm run dev          # all five editors + shell against Vite dev servers
npm run dev:docs     # a single app (same pattern works per workspace)
npm run dist:mac     # package macOS dmg (regenerates third-party notices)
npm run dist:win     # package Windows nsis installer
npm run dist:linux   # package Linux AppImage + deb + rpm

The sheets app additionally needs a Rust toolchain for its xlsx sidecar (cargo on PATH); npm run build -w @genoffice/sheets compiles it automatically.

Local UI/e2e driver scripts (Playwright + Electron, for local acceptance, not committed by default) live in scripts/drivers/.

Architecture notes (docx round trip)

open docx ─► archive original by hash (never touched)
          ─► docx-engine parses word/document.xml top-level elements (w:p / w:tbl / …)
          ─► Block tree, each block anchored by docxIndex + original XML slice
          ─► Tiptap streaming editor (manual + AI editing, dirty tracking)
save      ─► dirty blocks → OOXML fragments (referencing existing styles only)
          ─► splice into original document.xml (untouched blocks keep original bytes)
          ─► repack zip; all other entries copied byte-for-byte

The same philosophy holds in sheets and slides: the original file is the source of truth, edits are applied as narrow patches, and everything the editor didn't touch survives the round trip untouched.

FAQ

Is GenOffice free? Yes. GenOffice is free and open-source under the Apache-2.0 license — no trial, no paid tier for the apps themselves.

Can GenOffice open Microsoft Word, Excel, and PowerPoint files? Yes. GenOffice opens and saves native .docx, .xlsx, and .pptx files. Saving is byte-preserving: parts of the file you didn't touch are written back byte-for-byte, so documents keep working in Microsoft Office.

Does GenOffice work offline? Document editing is fully local — files never leave your machine to be opened, edited, or saved. The AI features need a network connection to whatever endpoint you configure in Settings → AI Provider — a LAN-only local server (e.g. Ollama) keeps that traffic off the public internet too; a cloud provider (Genspark, OpenRouter, a named vendor) does not.

Can GenOffice edit PDF files? Yes — real PDF text and image editing that rewrites the page content stream with the original fonts preserved, not cover-up annotations.

Security

See SECURITY.md for the process security posture (renderer sandboxing, IPC validation, external-link gating) and the threat models for AI-generated content.

Acknowledgements

GenOffice would not be possible without these open-source projects:

Electron — the desktop runtime for every app.
Univer (Apache-2.0) — the spreadsheet UI core that Sheets extends.
PDFium (BSD-3-Clause, bundled via @embedpdf/pdfium) — the content-stream engine behind true PDF text and image editing.
pdf.js (Apache-2.0) and pdf-lib (MIT) — PDF rendering and document assembly.
Tiptap / ProseMirror — the block editors in Docs and Markdown.
Konva — canvas rendering for Slides and Sheets charts.
HarfBuzz (wasm) — text-shaping metrics for complex scripts.
calamine and IronCalc — the read and calc layers of the Rust xlsx sidecar.
Liberation, Carlito, Caladea, and Noto CJK fonts (OFL/Apache-2.0) — bundled document fonts.

Third-party notices

npm run notices regenerates the bundled third-party license summary (tools/gen-third-party-notices.mjs); all runtime dependencies are MIT/Apache-2.0/BSD-3-Clause/OFL, and the bundled fonts (Liberation, Carlito, Caladea, Noto CJK subsets) are OFL/Apache.

License

GenOffice is licensed under the Apache License 2.0, with one exception: the ee/ directory is reserved for future enterprise modules and is covered by the GenOffice Enterprise License.

The GenOffice and Genspark names and logos are trademarks of Mainfunc, Inc. The Apache-2.0 license does not grant permission to use them (see section 6); forks should use their own branding.