Online Document Editor
An open source document workspace for you and your AI agents.
Open, edit, convert and save Word, Excel and PowerPoint files, and annotate PDFs in your browser. Core document processing runs locally, with no account or Office installation required.
Use optional AI assistance (Experimental Beta), let browser agents call document tools through WebMCP, or install it as a PWA. Cached editor resources work offline; you can also self-host or embed the editor in your app.
Try it online: edit.chaxus.com
✨ Features
- Edit locally. Open, edit and convert documents on your device, without a required document upload.
- Save your way. Write back to your chosen file where supported, or download an export. Reopen recent work from this browser.
- Work with AI (Experimental Beta). Draft and refine text, work with spreadsheet data, create slides or ask about the current PDF page. Review changes before applying them.
- Let agents handle documents. In WebMCP-compatible browsers, AI agents can open, convert and export files, and read text from word-processing documents.
- Install as a PWA. Use it as an app; cached editor resources can work offline.
- Host or embed it. Self-host a static site or add the editor to your app through its iframe API.
🚀 Quick start
New document · New spreadsheet · New presentation
Or run your own copy:
docker run -d --name document -p 8080:80 ghcr.io/ranuts/document:latest
📄 Formats
| Kind | Edit | Also opens |
|---|
| Documents | .docx | .doc .odt .rtf .txt |
| Spreadsheets | .xlsx .csv | .xls .ods |
| Presentations | .pptx | .ppt .odp |
| PDF | annotate, fill, export | .pdf |
Export supported documents to PDF.
Multilingual interfaces, light and dark themes, and read-only preview are also supported.
🔐 Local editing and data handling
Recovery copies stay in this browser for 7 days after the last edit or open. Delete them in Saved documents. Export anything you want to keep; recovery is not a backup.
Optional AI services receive the content needed for your request. On-device models keep it local. Browser dictation may use an online speech service; embedded exports are handled by the host app.
Your chosen browser AI agent controls how it handles content returned by document tools.
Documentation
User guide · Self-hosting, URLs and integration · Embedding API · Fonts
Offline use needs cached editor assets; uncached resources and remote files still need a network.
AI document assistant (Experimental Beta)
Draft and refine text, work with spreadsheet data, create slides or ask about the current PDF page. Review changes before applying and use Undo when needed.
Choose an on-device model or connect your own AI service. Dictate prompts where the browser supports speech input.
Experimental Beta. Off by default. Answers and edits can be wrong; check important results.
| Word | Excel |
|---|
| |
| Draft, write, summarize and undo | Fill, calculate, sort and undo |
| PowerPoint | PDF |
| |
| Create, duplicate and undo | Read, ask questions and review notes |
Complete workflows at normal speed, recorded in English with a loaded local Qwen3 4B model. Response times vary by model and device.
Videos, captions and text descriptions
PDF assistance covers selectable text and page notes, not scanned-image text or rewriting PDF body text.
This is the beginning of the README. Read the rest at the original. Read the full README ↗
