Skip to main content
Selective export lets you take control of what ends up in your exported file. Instead of exporting an entire conversation, you enter selection mode, check the individual messages you want, and export just those. The export button updates in real time to show how many messages you have selected, so you always know what you’re about to save.

​
How to use selective export

1

Open the ChatCache popup

Navigate to any ChatGPT conversation and click the ChatCache icon in your browser toolbar to open the popup.
2

Enter selection mode

Click the Select messages button in the popup. ChatCache activates selection mode on the page and checkboxes appear next to every message in the conversation.
3

Check the messages you want

Click the checkbox on any message to include it in your export. You can select your own prompts, the AI’s responses, or any combination across the thread.
4

Confirm your count and export

The export button in the popup updates in real time to show the number of messages selected (for example, Export 4 messages). When your selection looks right, choose your format and click the button to download or copy.

​
When to use selective export

Cherry-pick the best answers

Long research threads often contain a handful of genuinely useful responses buried among back-and-forth clarifications. Selective export lets you pull out only those answers without carrying along the surrounding noise.

Share only the relevant part

When you need to send a snippet of a conversation to a colleague or paste it into a document, select just the messages that are relevant to them rather than sharing the entire thread.

Export AI responses only

If you want a clean document of the AI’s output without your prompts, check only the assistant’s messages. This is useful for saving generated drafts, summaries, or code without the surrounding prompt context.
Use selective export to build focused knowledge bases. Pick the highest-quality answers from multiple conversations, export them to Markdown, and collect them in a folder in Obsidian or Notion for a curated reference you can search and reuse.