A single converter handles exports from ChatGPT, Claude, Gemini, Perplexity, Poe, Grok, Copilot, NotebookLM, and Google AI Studio. Request each platform's official export, drop the file in, and download conversations as Word, PDF, Markdown, CSV, or text without anything being uploaded.
Short answer
A single converter handles exports from ChatGPT, Claude, Gemini, Perplexity, Poe, Grok, Copilot, NotebookLM, and Google AI Studio. Request each platform's official export, drop the file in, and download conversations as Word, PDF, Markdown, CSV, or text without anything being uploaded.
Most people no longer use one assistant. Research happens in Perplexity, code in Claude, drafting in ChatGPT, and whatever Google puts in front of them for the rest. That leaves history scattered across accounts in incompatible formats, and one downloader per platform is a poor answer.
Where each platform keeps your history
| Platform | Where the export lives | Format |
| --- | --- | --- |
| ChatGPT | Settings, Data Controls, Export Data | conversations.json in a ZIP |
| Claude | Settings, Privacy, Export Data | conversations.json in a ZIP |
| Gemini | Google Takeout, Gemini Apps | JSON or HTML |
| Copilot | Microsoft privacy dashboard | JSON |
| Perplexity | Per-thread export and account data request | Markdown or JSON |
| Poe | Settings data export | JSON |
| Grok | X data request | JSON |
| NotebookLM | Per-notebook export via Google | Text and Docs |
| Google AI Studio | Prompt files in Drive | JSON |
The pattern is consistent: every vendor gives you machine data and none of them give you documents. The missing step is always the same one.
What a cross-platform downloader has to get right
Format detection rather than asking you which platform a file came from, since export shapes change without notice
Tree walking on ChatGPT and Claude, where edited and regenerated messages create branches
Takeout folder traversal for Gemini, where the data sits several directories deep
Unicode handling so emoji, Arabic, Chinese, and Cyrillic survive into PDF instead of becoming empty boxes
Stable naming so files from nine platforms sort into one coherent archive
One archive, one structure
The practical value is not the conversion itself, it is ending up with a single folder where a Claude thread from March and a Perplexity research session from July look the same and search the same. Once conversations are Word, PDF, Markdown, or CSV files on disk, your operating system search, Obsidian, Notion, or Excel handle them without knowing anything about AI vendors.
Local processing
ChatExports runs entirely in the browser tab. No account connection to any of the nine platforms, no OAuth grants, and no upload. Since these archives collectively contain most of what you have typed to a machine in the last two years, that is the only sensible default.
The All Platforms plan is a single one-time $25 payment covering every platform above, rather than paying per service.
Frequently asked questions
Do I need a separate tool for each platform?
No. One converter reads all nine export formats and detects which platform a file came from automatically.
Can I mix platforms in one session?
Yes. Drop several archives in and the conversation list combines them, so a single download can span platforms.
What happens when a platform changes its export format?
The parsers are updated for new shapes, and older archive formats keep working so previously saved exports still convert.
Is anything uploaded?
No. Files are read from disk and converted in the browser, then written back out as downloads.