GitHub - Oxeegen/OxeeOffice: Oxeegen's AI office suite — Word, Excel, PowerPoint, PDF, Markdown and HTML with an AI agent in every editor, running on Oxeegen's own models. Windows and Linux desktop app. · GitHub
Skip to content

About

Oxeegen's AI office suite — Word, Excel, PowerPoint, PDF, Markdown and HTML with an AI agent in every editor, running on Oxeegen's own models. Windows and Linux desktop app.

Resources

Code of conduct

Contributing

Security policy

Stars

3 stars

Watchers

0 watching

Forks

 
 

Latest commit

 

History

1,003 Commits

Folders and files

Repository files navigation

OxeeOffice

OxeeOffice

English · Français

Oxeegen's AI office suite.
Word, Excel, PowerPoint and PDF files, edited by you and your AI, saved back in the real formats — running on Oxeegen's own models.

License: Apache-2.0 Latest release Downloads Platform: Windows and Linux

Download · CLI · MCP · Oxeegen · Privacy

OxeeOffice is a desktop office suite for Windows and Linux. It opens and saves native .docx, .xlsx and .pptx files, edits PDF, Markdown and HTML, and puts an AI agent next to every document — not a chat box bolted on the side, but an editor that reads the file, makes the change, and shows you exactly what it touched.

  • Real formats, byte-preserving. Only what you edit is rewritten. Everything else in the file survives byte-for-byte, so documents keep working in Word, Excel and PowerPoint.
  • AI you can review. Edits land as tracked changes and diffs with one-click rollback. Spreadsheets get live formulas, not pasted numbers. Decks and pages are generated onto the canvas and stay fully editable.
  • Local by design. Files open, edit, save and convert on your machine. PDF → Word / Excel / PowerPoint, Markdown → Word and HTML → Word all run on-device. Only the AI calls leave the machine, to the provider you choose.
  • Oxeegen models by default. AI runs on Oxeegen's self-hosted, OpenAI-compatible inference. You can also bring your own key for Claude, OpenAI, Gemini, DeepSeek, Mistral, OpenRouter and others, or any OpenAI-compatible endpoint, local servers included.
  • Scriptable and agent-ready. The app ships a command line and a skill for Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, OpenCode and Windsurf, so a coding agent can create, convert, read and edit real Office files on your machine without opening a window.
  • Updates itself. Installed apps check this repository's releases and offer each new version.

Get it: Windows (x64) · Linux (deb, rpm, AppImage) — details and requirements in Download.

The apps

Six editors, one AI panel, and a command line for your coding agent.

1 · Docs — open and edit .docx with an AI you can review

Docs rendering a two-column annual-report page with a full-width cover image, shaded KPI table, header and footer, at 80% zoom with the AI panel collapsed Docs: a company overview with a banner image; the AI tightened the Overview and inserted a new bulleted section, and the panel offers a one-click roll back
Opens the file as Word lays it out — two-column sections, full-bleed images, shaded tables, headers and footers, pagination on Word's line metrics. Styles, comments, tracked changes, equations and ink round-trip untouched. Ask for the edit — the AI reads the blocks it needs, rewrites the Overview and inserts a new bulleted section. Every AI turn is a snapshot you can roll back; with Track changes on, edits arrive as Word-style revisions.

2 · Sheets — .xlsx with live formulas and charts, not pasted numbers

Sheets: the AI added a Summary sheet with revenue by region and category using SUMIF formulas, plus a column chart, and reports 43 applied changes with an Undo button Sheets: asked which region led Q2 revenue, the AI answers Europe with the category breakdown and cites the cells it used as links, next to the Orders sheet
Build it — from one sentence the agent adds a Summary sheet with real SUMIFs by region and category, inserts a column chart, and applies the 43 changes as a single undoable batch. Ask it — questions about the workbook come back with the reasoning and the exact cells used as clickable citations. Under the hood: a Rust .xlsx engine, pivot tables, slicers, conditional formatting and formula tracing.

3 · Slides — from a prompt to a .pptx deck

Time-lapse of Slides generating the Aurora Home investor deck: the AI plans the storyline in the panel, slides appear on the canvas one after another, and the finished deck ends on the closing ask

Slides: the cover slide of an AI-generated Aurora Home investor deck on the canvas, with the original one-line prompt and the AI's summary of what it built in the panel Slides: the designed closing slide of the same 11-slide deck, with the thumbnail strip on the left and the AI panel summarizing the storyline
One line in — "Create a 10-slide investor pitch deck for Aurora Home…". The AI plans the storyline, researches the numbers, and drafts every slide onto the canvas as a real .pptx. A finished deck out — designed slides with consistent typography, imagery and a closing call to action; keep editing with masters, layouts, smart guides and non-destructive cropping, or ask the panel to restyle, rewrite and reorder.

4 · PDF — edit PDF text in place, convert PDF to Word on-device

PDF: Edit text mode outlines every text block on the page for in-place editing while the AI panel answers a question about the report with page citations Docs showing a Word document converted locally from the Helios quarterly review PDF, opened in a second tab beside the original PDF
Edit inside the page — Edit text mode outlines every text block for in-place retyping; the content stream is rewritten through PDFium with the original fonts, not a cover-up annotation. Ask the AI about a long report and get answers with page citations. Convert on-device — PDF Converter → PDF to Word produces an editable .docx that opens in Docs next to the source, headings, stat rows and paragraphs intact. Excel and PowerPoint targets work the same way; scanned pages go through the system OCR.

5 · HTML — an AI page and UI builder, design brief first

Say what the page is for and who it is for. The AI proposes a design brief first — hook, palette, typography and style directions — then builds a single self-contained .html file against those tokens.

Time-lapse of HTML restyling the Lumen landing page: one Restyle request in the panel turns the dark Midnight Studio page into the warm Solar Daybreak version while every section and all copy stay in place

HTML: a generated landing page for a solar desk lamp in the dark Midnight Studio direction, shown in the live preview with the AI panel summarizing the page it just built The same Lumen landing page restyled by the AI into the warm Solar Daybreak direction: paper background, serif headlines and an orange accent, with every section and all copy kept
Generated from one prompt — a bold hero, feature cards, pricing and a waitlist form, built in the Midnight Studio direction. Click any element to restyle it, double-click to edit text, or switch to the CodeMirror source view. Same design, new direction — one Restyle request swaps the brief's tokens and the page follows: warm paper, editorial serif, sun-orange accent, nothing rewritten. Present fullscreen, or export as PDF or a native editable Word document.
HTML: a generated personal dashboard UI for a freelance designer in a warm linen style, with a left rail, serif greeting and four metric cards HTML: a generated EV-market data report in a broadsheet style, with a serif masthead, a 17.3 million headline figure and a stat row
UI mockups — the "personal dashboard" starter turns a persona into a working layout: left rail, greeting, billable-hours sparkline, invoice and utilization cards, all real HTML you can hand to a developer. Data stories — the "data report" starter builds an editorial broadsheet: serif masthead, one headline number, a rule-separated stat row, inline SVG charts and a methodology note.

6 · Markdown — a block editor over plain .md, with Ask AI

Markdown: a selected paragraph shows an Ask AI popover with a typed instruction and suggestion chips such as Polish, Make more concise, Expand and Fix grammar, plus Send now and Add to queue buttons Markdown rendering a launch-notes document with a table, a Mermaid flowchart and a task list, with the AI panel's starter prompts on the left
Ask AI about a selection — select any passage and an Ask AI chip appears: type an instruction or pick a suggestion, send it now, or queue several anchored edits and run them in one pass. The same entry exists in every app. Rendered, saved as plain Markdown — headings, lists, tables, images, code blocks and Mermaid diagrams in a Tiptap block editor, written back as plain .md, with a fully local Markdown → Word export.

7 · CLI — your coding agent drives OxeeOffice, on your machine

OxeeOffice ships a command line and an agent skill. Install the skill and Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, OpenCode or Windsurf can create, convert, read and edit real Office files through the same engines as the apps, without opening a window.

Slides showing an eight-slide Solar System deck that a coding agent built through the command line: the cover slide on the canvas, eight thumbnails on the left and the AI panel open

The eight rendered slides of the Solar System deck side by side: cover, exploration timeline, four key numbers, planet-diameter bar chart, rocky worlds versus giants, the Sun's 99.8% hero number, the four giants grid and takeaways Settings, Integrations page: the skill installed into Claude Code, with Install buttons next to Codex and Cursor
One prompt to your agent — "Build an eight-slide deck about the Solar System." The agent reads the skill, writes a style sheet, an outline and one page spec per slide, and lets slides check reject anything that overflows or overlaps before create assembles the .pptx and slides render hands back a PNG per slide to look at. Install once, from Settings → Integrations — OxeeOffice lists the coding agents it finds on this computer and writes the skill into each one you pick. Or download the skill as a zip, or run npx skills add Oxeegen/OxeeOffice. See Command line and agent skill.

8 · MCP — the same tools over the Model Context Protocol

Every command is also an MCP tool. Claude Code, Claude Desktop, Cursor and any other MCP client can start the server themselves, with no skill to install and no window open. A second, HTTP server inside the app lets an agent build a Word document in a visible editor tab while you watch.

Time-lapse of Claude Code building an eight-slide renewable-energy investor briefing through the MCP server: it searches for figures and photos, checks each candidate picture, writes the style sheet and outline, adds one checked page at a time, assembles the .pptx and renders every slide; the finished deck then opens in Slides

Slides showing the eight-slide Renewable Energy 2026 deck that Claude Code built through the MCP server: the cover slide with a wind-farm photograph on the canvas and eight thumbnails on the left Settings, Integrations page, MCP part: the one-line claude mcp add command for Claude Code, the JSON block for Cursor, Claude Desktop and other MCP clients, and the local HTTP server option below
One prompt, no shell — "Build an eight-slide investor briefing about renewable energy in 2026, with a real photo on the cover." The agent pulls figures and photos with search, checks each picture with media, builds the deck page by page with every page checked against the outline and palette, then renders each slide as an image it can look at and fixes the pages it does not like. Connect once, from Settings → Integrations — copy the claude mcp add line for Claude Code, or the JSON block into Cursor, Claude Desktop or any other MCP client. See MCP server.

Why OxeeOffice

  • Runs on Oxeegen. AI on Oxeegen's own inference, with the region endpoint configurable per user — no third-party account or subscription.
  • Yours to run. Native apps for Windows and Linux; files stay on your disk and every edit, save and conversion happens on your machine.
  • Real Office files. Native .docx, .xlsx and .pptx, byte-preserving: the parts of a file you did not touch are copied exactly as they were.
  • An AI that edits the document itself. Tracked changes in Docs, live formulas and charts in Sheets, slides drawn onto the canvas, every AI turn a snapshot you can roll back.
  • PDF done properly. Edit text inside the page, and convert PDF to Word, Excel or PowerPoint on-device, with system OCR for scans.
  • Markdown and HTML too, with the same AI panel and local export to Word.
  • Scriptable. A command line, an agent skill and an MCP server put every engine at the service of your coding agents, still on-device.
  • Open source, Apache-2.0, and free.

AI backends

Oxeegen is the default provider: chat, image and video analysis run on Oxeegen's models (Max, Pro, Flash, Instant) through its OpenAI-compatible endpoint. In Settings → AI Model, pick your region (US or EU — each has its own keys) and paste your Oxeegen API key; the endpoint is filled in for you. Switch models any time from the picker at the top of each editor's AI panel: that model plans with reasoning on, then writes slides, checks slide layout and writes long documents with reasoning off. Keys are never bundled, because releases are public.

Under AI Media & Search, web and image search run on Brave Search — paste a Brave API key into the Oxeegen search entry — and image generation uses OpenAI (gpt-image-2.5-flare, medium quality) with your OpenAI key, until Oxeegen serves its own image models.

Or bring your own key. Settings → AI Model also lists Claude, OpenAI, Gemini, DeepSeek, Kimi, GLM, Qwen, Doubao, MiniMax, Grok, Mistral, OpenRouter, Requesty and OpenCode Zen/Go, plus a custom slot for any OpenAI-compatible endpoint (base URL + key), including local model servers. Search and media have their own per-capability providers under AI Media & Search.

The whole suite ships light, dark and system themes. Themes only change what is on screen: exports, prints and saved files always keep the document's own colors.

Command line and agent skill

Everything the apps can do to a file, the command line can do from a terminal: inspect, convert, create, read and edit Word, Excel, PowerPoint, PDF, Markdown and HTML on the same engines, headless. It installs with OxeeOffice, needs no runtime of its own, and never sends a document anywhere. Paired with the bundled agent skill, it turns a coding agent into a document worker that produces real Office files instead of Markdown approximations.

The command is still called genoffice in this release; it becomes oxeeoffice in an upcoming one, together with the MCP server name and the skill. The examples below use the current name.

Works with: Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, OpenCode and Windsurf out of the box, any other agent that reads skills, and, through the MCP server, Claude Desktop and every MCP client.

Install the skill

How What happens
Settings → Integrations in the app Lists the agents found on this computer; one click writes the skill into each one you choose. An Update button appears when a release ships a newer skill.
Download as zip on the same page The layout claude.ai, the Claude desktop apps and other assistants accept as an uploaded skill.
npx skills add Oxeegen/OxeeOffice Installs from this repository into any skills-compatible agent.

Then start a new chat and ask for a document. The skill teaches the agent when to reach for the command line, how to read a file before editing it, and how to check its own work.

Quickstart from the terminal

genoffice --version
genoffice info report.docx --json                  # headings and blocks; or sheets, slides, pages
genoffice convert report.md --to pdf               # md/html/docx/xlsx/pptx → pdf, pdf → docx/xlsx/pptx, …
genoffice create --type docx --from notes.md --out notes.docx
genoffice create --type xlsx --from table.json --out sales.xlsx   # "=SUM(B2:B9)" cells stay live formulas
genoffice docs read report.docx --range 0-9 --json # then `docs apply --ops edits.json` edits in place
genoffice render report.docx --out shots/          # one PNG per page, to look at what you made
genoffice open sales.xlsx                          # hand the result to the editor

Every command prints a one-line summary, or a single JSON object with --json. Edits are atomic: a rejected op leaves the file untouched and comes back with a guided error. genoffice help lists the current command surface; the full reference is packages/cli/README.md.

Building a deck from an agent

An agent following the skill's staged workflow runs something like this, and the command line checks every stage before the next one starts:

genoffice capabilities --json                        # which cloud tools are configured
genoffice guide slides design                        # the deck workflow and layout library
genoffice slides check deck/outline.json --json      # outline checked
genoffice slides check deck/pages/01.json --json     # builds one slide, audits overflow and overlap
…                                                    # one page file per slide, fixed until each check is clean
genoffice create --type pptx --spec deck/pages --outline deck/outline.json --out deck/briefing.pptx --json
genoffice slides render deck/briefing.pptx --out deck/shots --json
genoffice slides audit deck/briefing.pptx --json
genoffice open deck/briefing.pptx

No model call happens inside the command line: the agent does the thinking, the command line does the building and the checking, and the result opens in OxeeOffice or PowerPoint as an ordinary .pptx.

MCP server

The same commands are available as Model Context Protocol tools, for assistants that cannot run a terminal or that you would rather not give one. There are two ways in, both shown with copy-ready snippets in Settings → Integrations → MCP:

Way What it is
A · genoffice mcp (recommended) A stdio server the assistant starts itself; OxeeOffice does not need to be open. One tool per command (info, convert, create_*, docs_*, sheet_*, slides_*, render, guide, search, image, media, open) plus the staged deck flow deck_start → deck_page → deck_build → deck_replace.
B · Local HTTP server Runs inside the app on http://127.0.0.1:3093/mcp. Its tools drive a visible Word editor tab, so you watch the document take shape. Off by default; switch it on in the same settings pane.
# Claude Code
claude mcp add --transport stdio genoffice -- genoffice mcp
// Cursor, Claude Desktop or any other MCP client
{ "mcpServers": { "genoffice": { "command": "genoffice", "args": ["mcp"] } } }

genoffice here is the command line shipped inside the app — on Windows %LOCALAPPDATA%\Programs\OxeeOffice\resources\cli\genoffice.cmd, on Linux /usr/bin/genoffice; the settings pane prints the exact path for your install. Cloud features (search, image, media) go through the provider configured in OxeeOffice; everything else runs locally, and GENOFFICE_ALLOWED_ROOTS confines every tool to the folders you list.

Download

Platform Requirements Download
Windows (x64) Windows 10+, Intel/AMD OxeeOffice-<version>-setup.exe
Linux — Debian / Ubuntu x86_64, glibc 2.34+ (Ubuntu 22.04 or newer) oxeeoffice_<version>_amd64.deb
Linux — Fedora / RHEL / openSUSE x86_64, glibc 2.34+ (Fedora 35+, RHEL 9+, Leap 15.6+) oxeeoffice-<version>.x86_64.rpm
Linux — other distributions x86_64, glibc 2.34+, FUSE 2 OxeeOffice-<version>.AppImage

The Windows installer and the AppImage update themselves; deb and rpm installs update through the package manager. The installers are unsigned: on first run, Windows SmartScreen warns — choose More info → Run anyway. macOS is not built, because without an Apple Developer certificate Gatekeeper refuses the app.

What changed in each version is in CHANGELOG.md.

Installing on Linux

The deb installs with apt — it pulls in the dependencies and adds OxeeOffice to the applications menu:

sudo apt install ./oxeeoffice_<version>_amd64.deb

On Fedora / RHEL-family / openSUSE, install the rpm instead:

sudo dnf install ./oxeeoffice-<version>.x86_64.rpm     # Fedora / RHEL family
sudo zypper install ./oxeeoffice-<version>.x86_64.rpm  # openSUSE

The AppImage runs in place: install the FUSE 2 runtime (sudo apt install libfuse2; on Ubuntu 24.04 the package is libfuse2t64), make the file executable, then run it:

chmod +x OxeeOffice-<version>.AppImage
./OxeeOffice-<version>.AppImage

How it works

Seven Electron apps — Docs, Sheets, Slides, PDF, Markdown, HTML and the tabbed shell — share one engine layer of pure TypeScript packages plus a Rust sidecar for .xlsx. The original file is always the source of truth: edits are applied as narrow patches, and everything the editor did not touch survives the round trip untouched.

open docx ─► archive original by hash (never touched)
          ─► parse word/document.xml into a block tree, each block anchored to its original XML
          ─► Tiptap editor (manual + AI editing, dirty tracking)
save      ─► dirty blocks → OOXML fragments (referencing existing styles only)
          ─► splice into the original document.xml; untouched blocks keep their bytes
          ─► repack the zip; every other entry is copied byte-for-byte

Development

npm install
npm test             # engine + app unit tests
npm run typecheck    # tsc --noEmit across every workspace
npm run dev          # all six editors + shell against Vite dev servers
npm run dev:docs     # a single app (same pattern works per workspace)

The Sheets app needs a Rust toolchain for its xlsx sidecar (cargo on PATH). Installers are built by the release workflow. Setup, checks, releases and pull requests are in CONTRIBUTING.md.

Support

FAQ

Is OxeeOffice free?

Yes. The apps are free and open source under the Apache-2.0 license. AI features use Oxeegen inference or the provider key you configure.

Can OxeeOffice open Microsoft Word, Excel and PowerPoint files?

Yes. OxeeOffice 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 OxeeOffice work offline?

Document editing is fully local — files never leave your machine to be opened, edited, saved or converted. The AI features need a network connection to Oxeegen or to the provider you configured.

Can OxeeOffice edit PDF files, or convert them to Word?

Yes to both: real PDF text and image editing that rewrites the page content stream with the original fonts, and PDF → Word / Excel / PowerPoint conversion entirely on-device, with system OCR for scanned pages.

Can I use a different AI model or my own API key?

Yes. Oxeegen is the default, and Settings also accepts keys for Claude, OpenAI, Gemini, DeepSeek and others, or any OpenAI-compatible endpoint, including local model servers.

Can I drive OxeeOffice from Claude Code, Codex, Cursor or a script?

Yes, through the bundled command line and agent skill, or the MCP server. See Command line and agent skill.

Does OxeeOffice collect any data?

No. OxeeOffice builds send no usage analytics: they are never given analytics credentials, and the release workflow fails if a package carries them. Documents stay on your machine; only the AI requests you make are sent, to the provider you configured. See PRIVACY.md.

Security

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

License

OxeeOffice is open-source software under the Apache License 2.0. Attribution and third-party notices are in NOTICE.

The ee/ directory is covered by a separate license (ee/LICENSE) and is not part of OxeeOffice builds.

Maintained by Oxeegen.

About

Oxeegen's AI office suite — Word, Excel, PowerPoint, PDF, Markdown and HTML with an AI agent in every editor, running on Oxeegen's own models. Windows and Linux desktop app.

Resources

Code of conduct

Contributing

Security policy

Stars

3 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages