Export full conversation histories from your favourite AI assistants into clean files — no account, no backend, nothing leaves your browser.
Save the active conversation as a clean file without sending content to a server.
Most exporters break because they rely on a single set of fragile selectors. ChatArchive uses per-platform adapters with multiple fallback strategies.
API → page state → DOM scroll, with automatic fallback. Prefers stable data sources when the platform exposes them.
Viewport-based scrolling with batched DOM collection. Handles lazy-loaded conversations on all supported platforms.
See exactly how each extraction worked — source, strategy, confidence score, and message count, right in the popup.
Use tokens like {platform}, {title}, {date} to keep your archive organised automatically.
Press Ctrl+Shift+S at any point to save what's been captured so far — useful for very long conversations.
Fixture-based extraction tests with jsdom, plus scheduled live Playwright smoke checks to keep platform support visible.
Format preference, filename template, message limit, and scroll speed are all saved across browser sessions.
Skip the file download entirely — copy the formatted conversation straight to your clipboard with one click.
No configuration needed. Open any supported chat, click the extension.
Navigate to any supported AI assistant with a conversation you want to save.
ChatArchive detects the platform, scrolls through the full history, and extracts every message.
Your conversation downloads as a clean file in whichever format you chose. Nothing is uploaded.
From Obsidian vaults to data pipelines — pick the format that fits your workflow.
Each platform has its own dedicated adapter with multiple extraction strategies and automatic fallback.
ChatArchive has no server. Every byte of processing happens inside your browser.
There is no server. Exports are generated entirely within your browser's local context.
Nothing is uploaded, logged, or shared. Your conversations never leave your machine.
The full source code is on GitHub. You can read, audit, or fork every line.
No analytics, no usage reporting, no advertising. The extension does one thing.