Frequently asked questions

It depends on the format. Word, Markdown, JSON, TXT and long screenshots are generated locally in your browser and are never sent anywhere. PDF is rendered on our server by default, so your conversation is uploaded for rendering — you can switch PDF to a local engine in Settings if you want nothing to leave your browser. Optional sign-in and anonymous usage analytics are the only other network calls.

No. We don't train anything, and we don't have a data warehouse. There's no place for your conversations to land. The only thing we persist tied to a signed-in account is a daily export count, used only for capacity planning — never your chat content.

No. PDFs export without any watermark.

No. It only activates on the 14 AI domains it's whitelisted for. Open the manifest in chrome://extensions — the host_permissions field lists every site it can read, and that's the full list.

Fourteen: ChatGPT, Claude, Gemini, Google AI Studio, NotebookLM, DeepSeek, Grok, Microsoft Copilot, Perplexity, Kimi, Tencent Yuanbao, Doubao, Qwen, and GitHub conversations. We add more once they're stable.

If a conversation is very long, some platforms lazy-load older messages and don't keep them in the DOM. Scroll to the top of the chat before exporting — that forces the page to re-render the full transcript.

Yes, in PDF and long-screenshot formats. Markdown export keeps the language tag (```python etc.) so your downstream tool can highlight it. Word turns code blocks into monospace boxes.

All preserved in PDF and long-screenshot. Markdown export uses standard image syntax and LaTeX-style math. JSON export gives you raw text plus image URLs for downstream pipelines.

Not yet — current scope is one conversation per click. Batch export is on the roadmap. If you need it sooner, send a note and we'll bump it.

Open the conversation, click the Save AI toolbar icon (or the camera button that appears on a message), and pick a format. The file downloads immediately — no setup required.

If you just installed the extension, refresh the AI tab once so the content script loads. The in-page button appears only on supported sites, and on some of them only when you hover over a message's action row. The toolbar icon and the right-click menu always work.

Images hosted on another domain need permission before they can be embedded. The first time you export images from a new site, accept the permission prompt. Images from ChatGPT, Gemini and Yuanbao are included automatically.

Yes. Open the preview, then select individual messages, or use the menu to keep only the questions or only the answers, before you save.

Yes. The preview toolbar has theme and font size controls, and Settings offers eight document themes plus width, table style, font and PDF margin options.

It runs on Chrome today, with Edge and Firefox builds on the way. Browser extensions don't run on mobile browsers, so it is desktop-only for now.

Yes — everything is free and unlimited. PDF, Word, Markdown, JSON, TXT and long screenshots, with no caps and no watermark.

No. No ads, no affiliate links, no sponsored content. The extension does what it says on the tin and gets out of the way.

No. The extension works without signing in. Google sign-in is optional and currently only used for the small admin console.

Google sign-in only, for now. The reason is simple: we don't want to store passwords. Email/password is on the roadmap if there's demand.

Yes. The extension is per-browser; install it on each browser you use. There's no account-level state to sync right now.

Submit a request via our Feedback page (include the email tied to your account) and we'll wipe it. Account data is just an email, Google ID, and timestamps — nothing else.