ConvoArk
English

Guide

Install, first save, choosing what and how to save, and what to do if something goes wrong.

1. Install

  1. You need Google Chrome 141 or later on a computer (check at chrome://version).
  2. Open ConvoArk in the Chrome Web Store and click “Add to Chrome”.
  3. Chrome shows the permissions it needs; the FAQ explains each one.

2. Pin the toolbar icon

Click the puzzle-piece icon at the top right of Chrome, find ConvoArk and click the pin. The icon now stays in the toolbar, always within reach.

Saving always starts from this icon: Chrome lets the extension read a page only after you click it.

3. Save a conversation

  1. Open a conversation on ChatGPT, Claude or Gemini and wait until the reply has finished.
  2. Click the ConvoArk icon in the toolbar. Under “Format”, pick HTML, Markdown or PDF.
  3. Click “Save as offline file” (for Markdown, “Save as Markdown”; for PDF, “Save as PDF”). To keep only some messages, click “Choose messages to save…” instead (see the next section).
  4. The page reloads once so collapsed content can be expanded. Unsent text in the message box is lost, so send or copy it first.
  5. Progress shows on the page while saving. Keep that tab open and in front until it finishes; long conversations take longer.
  6. When it is done, the popup lists how many messages, images and citations were saved. A row marked in amber says what is missing, why, and how to complete it on the next save.

Gemini interactive modules: Chrome asks once whether ConvoArk may read them. Allow it and they stay clickable offline; choose “Save without access” and they are saved as a picture or text. This is the only extra permission prompt.

4. Choose messages before saving

5. Export Markdown

Choose Markdown under “Format” and click “Save as Markdown” (in the panel, the save button). The extension makes one .md file on its own print page and saves it where Chrome saves downloads. There is no image folder, so some content is written differently:

Some Markdown readers load image links when they open a file, which contacts the site hosting the image. To keep images and interactive content inside the file, choose HTML.

6. Export PDF

  1. Choose PDF under “Format” and click “Save as PDF” (in the panel, the save button). In the panel you can also tick “Start each round on a new page”.
  2. The extension opens its own print page in a new tab, lays out the conversation and opens Chrome’s print window.
  3. In the print window, set Destination to “Save as PDF”. Turn off headers and footers in the print settings (otherwise Chrome adds its own address and date), then save. If you closed the print window, click “Print again” on the print page.

The PDF is made locally in your browser. Nothing is uploaded and no extra permission is needed. Interactive modules appear as the picture taken when saving.

7. Where the file goes

HTML and Markdown files are saved wherever Chrome saves downloads: the Downloads folder unless you changed it, or the place you pick if Chrome asks. The name includes the conversation title. “Open in Downloads” in the popup takes you to it, and chrome://downloads lists it too. A PDF is saved where you choose in the print dialog.

8. Open it offline

9. Change the interface language

The popup and the selection panel come in Japanese, English, Simplified Chinese and Traditional Chinese. By default they follow your browser’s language: Traditional Chinese locales (such as zh-TW and zh-HK) show Traditional Chinese, any other Chinese shows Simplified Chinese, Japanese shows Japanese, and anything else shows English.

To change it, click the language button with the globe icon at the top right of the popup and pick from the “Language” menu. Choose “Automatic” to follow the browser again. Only buttons and messages change; conversations are never translated. A saved HTML file has its own language button (see above).

10. Temporary chats and pages without a conversation

11. Troubleshooting

The popup says “This isn’t a chat page”

Saving works only on a conversation page of chatgpt.com, claude.ai or gemini.google.com. Open the conversation itself (not the home page or settings) and click the icon again.

A reply is still being written

Wait for the reply to finish, or schedule the save in the popup so it starts when the reply is done.

The “Messages” row is marked in amber

Something didn’t add up, for example the start of a very long conversation didn’t finish loading. Let the page load fully and save again. If the warning stays, send us a screenshot of that row (no conversation content needed).

Some images show as placeholders

The picture could not be read or captured. The placeholder says why and links to the original. If it says the tab was not in front, save again and keep the conversation tab in front.

A Gemini module is only a picture

Module access was not allowed, or its code couldn’t be read. Click “Allow interactive modules” in the popup and save again.

I can’t find the file

Open chrome://downloads: the latest entry is the file, with a link to show it in its folder. A PDF is where you chose in the print dialog.

The icon doesn’t appear or nothing happens

Check that Chrome is version 141 or later and that ConvoArk is turned on at chrome://extensions. Then reload the conversation page and try again.

Still stuck? See the FAQ or Support.