All releases
ImprovedPrivacy

AI and your data, in one place

A new docs page says which AI powers Markdrop's AI features, what is sent to it and what never is, and what Markdrop keeps. AI features describe their limits as fair-use limits.

  • Web app
  • MCP server
  • API
  • AI and your data. Markdrop's AI features are powered by Anthropic's Claude models. Only the excerpts a request needs are sent; encrypted items never are, a password-protected item only by its title, and nothing sent is used to train models.
  • Fair-use limits. When an AI limit is reached, the feature keeps working without the AI part: meetings still import with their transcript and Google's notes and Write it again adds the AI sections later, Jira drafts are the item as written, and Team picks up again by itself.
  • Same in the API and MCP. A bulk meeting import's review says how many meetings get an AI summary and whether that's within today's fair-use limit, with no amounts; AI responses and tool results describe limits the same way.

Docs updated