# Transfer Any File

> Transfer Any File is an open-source Chrome extension (Manifest V3, MIT licence) that converts files between 14 formats — Markdown, HTML, Word (.docx), PDF, TXT, CSV, Excel (.xlsx), JSON, PNG, JPEG, WebP, BMP, GIF, SVG — entirely on the user's own machine. It requests one permission (`storage`), its own code issues no network request of any kind (asserted on every push by `pnpm verify:offline`), and it works with the network disabled. Batch conversion of mixed source formats, automatic multi-step conversion chains over 48 direct routes, per-file error isolation with copyable failure diagnostics, split preview with inline editing, a recently-used group in the target picker, ZIP download, and a local conversion history.

Primary language: the product page is bilingual with an in-page Chinese / English switch; the privacy policy carries both languages on one page. Neither needs a server. Interface languages: English and Chinese; with no stored choice the extension follows the browser language (a `zh*` browser gets Chinese, anything else English), and an explicit choice is remembered.
Category: productivity / file conversion. Price: free, no paid tier, no ads, no account.

## Product page

- [Product page](https://liaolongdong.github.io/transfer-any-file/): Formats, how a conversion is routed, features, comparison with online converters, install steps, FAQ, technical details, contact
- [Privacy policy](https://liaolongdong.github.io/transfer-any-file/privacy.html): Bilingual (EN / 简体中文) — no data collection, no transmission, on-device storage only

## What it converts

- Documents: Markdown, HTML, Word (.docx), PDF, TXT — in any direction, through HTML as the hub format
- Data: CSV, Excel (.xlsx), JSON — in any direction, including CSV ⇄ XLSX and JSON ⇄ CSV
- Images: PNG, JPEG, WebP, BMP, GIF, SVG — readable; writable as PNG, JPEG or WebP (GIF uses its first frame, SVG is flattened)
- Multi-step routes are discovered automatically: Markdown → HTML → PDF, Word → HTML → Markdown, JSON → CSV → XLSX, TXT → HTML → JSON → CSV → Excel
- 48 registered direct routes; breadth-first search over the same graph makes 143 source-to-target combinations reachable, and the format picker offers 116 of them. The other 27 are shown greyed-out with a reason rather than hidden (116 = 143 minus the 24 image→text/data and 3 PDF→data pairs listed below). Every source format reaches all 11 writable formats except itself.

## What it does not do (frequently asked of converters)

- No OCR: images cannot become TXT, CSV, JSON or XLSX (an OCR model would break the offline guarantee)
- PDF output is image-based, so text in a converted PDF is not selectable
- PDF input extracts text only; layout and embedded images are not preserved
- BMP, GIF and SVG are input-only — browsers expose no encoders for them
- No video, audio, e-book or CAD formats
- Limits: 100 MB per file (warned at 20 MB), 200 files per batch

## How it compares

| Dimension                         | Transfer Any File                                                             | Online converter                            | Command-line tool                                                  |
| --------------------------------- | ----------------------------------------------------------------------------- | ------------------------------------------- | ------------------------------------------------------------------ |
| File leaves the device            | Never                                                                         | Always (that is the mechanism)              | Never                                                              |
| Setup cost                        | Build once, load the unpacked folder                                          | None, it is a web page                      | Package manager and PATH                                           |
| Free-tier limits                  | 100 MB per file, 200 files per batch                                          | Daily quotas, size caps, watermarks         | None                                                               |
| Works offline                     | Yes                                                                           | No                                          | Yes                                                                |
| Batch of mixed formats            | Yes, one target for the whole batch                                           | Usually paid or queue-bound                 | Yes, and scriptable                                                |
| Preview and edit the result       | Built-in comparison view and inline editing                                   | Usually a thumbnail at best                 | None                                                               |
| Format breadth                    | 14 document / data / image formats                                            | Hundreds, including video and audio         | Pandoc 40+ document formats, ImageMagick hundreds of image formats |
| OCR and complex PDF layout repair | Not bundled                                                                   | Commonly offered                            | Requires a separate Tesseract install                              |
| Price and account                 | Free, no sign-up, no paid tier                                                | Free quota, a subscription lifts the limits | Usually free                                                       |
| Chinese interface and encoding    | Bilingual UI; CSV written with a UTF-8 BOM                                    | Depends on the service                      | Encoding is left to you                                            |
| Source code                       | Open source (MIT)                                                             | Closed                                      | Open source                                                        |

The comparison summarises the typical behaviour of each class of tool rather than one specific vendor; Pandoc and ImageMagick stand in for the command-line column. The three are not substitutes.

## How to install

Not published to the Chrome Web Store yet. Build from source (Node.js ≥ 20.12, pnpm), then load unpacked:

```bash
git clone https://github.com/liaolongdong/transfer-any-file
cd transfer-any-file
pnpm install && pnpm build
# chrome://extensions → enable Developer mode → Load unpacked → select .output/chrome-mv3
```

## Who maintains it

Sole author: Better ([github.com/liaolongdong](https://github.com/liaolongdong)). Contact channels, in the order they get answered: GitHub issues, email `924902324@qq.com` (also the address published on the store listing), and a WeChat group reached by adding `lld_1025` with the note `taf`. Security issues are reported privately, never as a public issue — see the security policy below. The same author maintains [Account Password Helper](https://github.com/liaolongdong/account-password-helper), a local-first password manager with the same no-cloud constraint.

## Repository

- [README (简体中文)](https://github.com/liaolongdong/transfer-any-file/blob/main/README.md)
- [README (English)](https://github.com/liaolongdong/transfer-any-file/blob/main/README.en.md)
- [Source code](https://github.com/liaolongdong/transfer-any-file)
- [Issues / feature requests](https://github.com/liaolongdong/transfer-any-file/issues)
- [Changelog](https://github.com/liaolongdong/transfer-any-file/blob/main/CHANGELOG.en.md): release notes per version ([简体中文](https://github.com/liaolongdong/transfer-any-file/blob/main/CHANGELOG.md))
- [Security policy](https://github.com/liaolongdong/transfer-any-file/blob/main/SECURITY.en.md): how to report a vulnerability, and the offline attack-surface claims you can verify ([简体中文](https://github.com/liaolongdong/transfer-any-file/blob/main/SECURITY.md))
- [Contributing guide](https://github.com/liaolongdong/transfer-any-file/blob/main/CONTRIBUTING.en.md): the offline-only rule, how to add a converter, and which docs must move together ([简体中文](https://github.com/liaolongdong/transfer-any-file/blob/main/CONTRIBUTING.md))
- [Chrome Web Store listing copy and disclosures](https://github.com/liaolongdong/transfer-any-file/blob/main/CHROMEWEBSTORE.md)

## Tech stack

WXT · Vue 3 · TypeScript · Element Plus · Vite · Playwright (end-to-end). Heavy conversion libraries (marked, turndown, mammoth, pdf.js, SheetJS, jsPDF, html-to-image, DOMPurify, fflate) are dynamically imported per converter; the complete bundle is 3.73 MB.