OpenAI blocks chat exports for Business users; free tool works around it

ChatGPT Business and Enterprise users have discovered they cannot export their workspace chats using OpenAI's standard data export feature, a limitation that affects only the highest-paying subscription tiers. OpenAI's help documentation explicitly states that Free, Plus, Pro, and some Edu customers can export their chats, but Business and Enterprise workspaces are excluded. The company has not disclosed why.

Faced with this restriction, freelance journalist Conrad Quilty-Harper built and released scrapemychats, a free tool published on GitHub this week. Quilty-Harper explained to The Register that he made the tool after upgrading from a personal ChatGPT account to a Business account—at OpenAI's prompting—only to find he could not easily retrieve his data. "I made the stupid decision to upgrade from a personal account to a Business one a while back as they prompted, and for some reason OpenAI doesn't let you export the data easily," he said.

OpenAI does offer Enterprise customers a path to preserve conversation logs through its Compliance Platform, which gives Enterprise admins access to workspace logs. However, that platform retains logs for only 30 days unless customers manually archive them elsewhere—a burden that falls short of what Business customers get, which is no option at all. Quilty-Harper suspects OpenAI's intent is to lock in top-tier subscribers by withholding this feature. The Register asked OpenAI for an explanation but did not receive a response.

Scrapemychats works by accessing your ChatGPT account through a logged-in browser session and retrieving your conversation data and file attachments, storing them in a local HTML archive that can be opened offline in any browser. Because OpenAI's bot protection prevents headless operation, the tool must navigate the ChatGPT interface manually, a process that is slow due to OpenAI's request throttling. According to the GitHub README, downloading a 600-chat archive takes a few hours, and the process can be paused and resumed. The tool is installed via command line, with all necessary steps documented in the GitHub repository. Quilty-Harper notes that the stored chats can be used however the user likes.

However, the window to use scrapemychats may be short. Quilty-Harper predicted to The Register that "OpenAI will stop this from working shortly after you write about it," acknowledging that the tool likely breaches OpenAI's terms of service. He added a pointed observation about the irony: "their entire business model is based around scraping copyrighted material. This just allows you to get the content you wrote or pasted or created (and generated) back." Quilty-Harper told The Register he would welcome an explanation from OpenAI: "I would quite like to know the reason why ChatGPT doesn't make it easy for paying customers to get access to their data."

OpenAI's decision to block chat exports for its highest-paying subscription tiers—Business and Enterprise—stands out because the feature is available to lower-tier customers. The company offers Enterprise admins access to conversation logs through its Compliance Platform, but that system retains data for only 30 days by default, forcing customers to manually archive conversations elsewhere if they wish to keep them longer. This friction appears deliberate; Quilty-Harper's public explanation that OpenAI likely aims to lock in premium subscribers aligns with the pattern of restricting a basic data-ownership feature to higher tiers.

The appearance of scrapemychats highlights a gap between what paying Business customers expect—the ability to own and port their data—and what OpenAI permits. The tool works within OpenAI's browser interface to sidestep the company's export restriction, storing conversations and attachments in an offline-readable format. However, Quilty-Harper's candid expectation that OpenAI will quickly patch the loophole underscores the tension: OpenAI has not publicly justified the export restriction, yet the company appears willing to defend it against workarounds. The irony Quilty-Harper noted—that OpenAI's business model relies on scraping copyrighted material, while the company restricts customers from exporting their own content—adds another layer to the friction.

No comments yet. Be the first to share your thoughts!