ChatExport VS Diffyn

ChatExport's answer

It does one thing: it turns a conversation into a document somebody else will accept. ChatExport reads a local iPhone backup, or the chat file WhatsApp exports itself, and writes a PDF, TXT, HTML, CSV or RSMF. The PDF is a real document rather than a database dump: title page, table of contents, PDF bookmarks by month, date separators, photos, attachments and read receipts. A second theme, Court View, drops the blue bubbles for a formal monochrome transcript. Every export is stamped with a SHA256 hash and gets a sidecar file next to it, and the document's own integrity page lists the checksums of the source files as they sit in the backup. The document language is chosen separately from the interface language, across fourteen languages, so you can work in your own language and file in another. Everything runs on your own Windows PC: no cloud, no account, nothing uploaded.

Diffyn's answer:

Addresses workflow and change management on LLM prompts, provide teams with traceability and visualization of tests across multiple models, provide deeper understading into efficiency of these prompts.

ChatExport's answer

Because of what comes out of it, and because of what it refuses to guess. Most alternatives fall into two groups: all-in-one phone managers that do thirty things and this one worst of all, and screenshot or OCR apps that photograph a screen instead of reading the record. ChatExport reads Apple's own database, including the attributedBody field that holds the text of many messages from iOS 16 onwards, where a tool reading only the plain text column returns messages with no words in them. It exports the whole thread or a date range, including the parts that do not help you, because selective screenshots are the first thing the other side attacks. It tells you when the history is incomplete, for example when Messages was set to delete after 30 days or when older messages may be sitting in iCloud rather than in the backup, instead of quietly handing you a shorter conversation. Encrypted backups are read with your password. The licence is one payment of 34.99 USD, never a subscription, with 30 days to get your money back, and there is a registered business behind it that you can look up before you pay.

Diffyn's answer:

Diffyn is the platform that specializes on both change management and multi-model analysis.

ChatExport's answer

A desktop app in Electron and TypeScript, with React in the interface and electron-vite for the build. Messages are read with better-sqlite3 from a read-only working copy of the backup's database, never from the backup itself, with a compatibility layer for the different layouts Apple shipped across iOS 11 to 26. Apple's binary property lists, including the attributedBody field, are decoded with plist and bplist-parser. The PDF is composed as HTML and CSS and printed by Electron's bundled Chromium, then assembled and stamped with pdf-lib and fontkit, which is what makes the output look like a document rather than a report generator's idea of one. Photos go through heic-convert and sharp, with the heavy work moved off the main process into worker threads. WhatsApp exports and the RSMF container are handled with yauzl and yazl, emoji are drawn from vector Twemoji bundled with the app, the interface is translated with i18next, and the tests run under Vitest. The installer is NSIS via electron-builder, signed with Azure Trusted Signing. There is no server side and no telemetry anywhere in it. The website is a separate, static Astro build.

Diffyn's answer:

React, Next.js, POSTGRESQL

ChatExport's answer

People who need a conversation in a form somebody else will accept, usually by a date that is already fixed. The largest group is parents in a custody or child support case, where the handover arrangements and the promises about money were all agreed by text and exist nowhere else. Alongside them: family and civil lawyers who would rather hand a judge a paginated transcript with a hash than a client folder full of screenshots, people in workplace and HR disputes, investigators and claims handlers who need the same backup to produce the same document every time, and journalists keeping a readable record of a source exchange on a machine that never uploads it. A smaller group has no deadline at all and is simply changing phones, and wants fifteen years of family messages in one PDF that will still open when the phone, the account and the app are long gone. In practice they are Windows users with an iPhone, or with a WhatsApp export from either phone.

Diffyn's answer:

Professionals incorporating LLMs or AI tools in their workflow and wants to keep track of changes and test their prompts.

ChatExport's answer

I wrote the first version for myself. After my divorce I was preparing for a child support and custody case, and almost everything that had been agreed over those months existed only as messages: dates, arrangements, who said they would do what and when. It had to go in front of a court, so I started taking screenshots. I took 150 of them and ended up with a folder of images carrying no dates, in no particular order, half of them cut off mid sentence. Any one of them falls apart under a single question: what was above it, and what came after? I looked at what else there was, and the programs that read an iPhone did thirty things at once and this one thing worst of all. I have written software for years, so I sat down and built it myself. The first PDF ChatExport ever produced was filed in my own case, and my lawyer was impressed by it. What a court accepts is always the court's decision, but the difference between a folder of screenshots and one ordered document is visible from the first page. Then it occurred to me how many people are in exactly the same position, so I finished it properly and put it out for everyone.

Diffyn's answer:

I started working on Diffyn when I notice that prompting has become an essential part of work across many industries. While there are version control platofrms like github, they are not designed for just prompt management are can be overkill such applications, it is also not integrated natively with various LLMs and relevant tools for users to validate ideas and visualise results properly.

ChatExport's answer

There are none to list, and I would rather say that than invent one. Two reasons, and both are deliberate. The licence went on sale on 25 August 2026, so the product is at the very beginning of its commercial life. More importantly, ChatExport contains no telemetry of any kind: it runs entirely offline, licences are verified on the machine without contacting a server, and nothing about who runs it, or what they export, ever reaches me. That is the core promise of the product, and knowing my users by name would mean breaking it. The people using it are individuals and small law practices working on family and employment cases, which is not the kind of work anyone wants their name attached to in a marketing list. When there are customers who are happy to be named, they will appear here with their permission and not before.