Appreciate MCP server availability, will help integrate with Claude. I literally have a Notion database where I tag papers by technique and link them to our codebase decisions. It's a nightmare to maintain. Does Paper Lantern keep any kind of project memory - like "for this repo, we previously decided on approach Y because of paper Z" - so the reasoning compounds over time?
yes - it instructs the coding agent to maintain a markdown file for the outputs from paper lantern, which the coding agent can refer to later. it also asks the coding agent to record whether the human user accepted the suggestion or not.
that way it helps you keep record of things - but also, you can come later and ask : which of our decisions now have updates in research ? can we adopt them ?