Tincture

A concentrated extract from a single paper. Read it deeply, connect it to the library, and distill actionable insights.

Arguments

The user provides a paper identifier (index number, title, arXiv ID, or citekey).

Steps

1. Full details: Call mcp__distillate__get_paper_details to get the complete paper record — summary, highlights, notes, metadata

2. Context in the library: Call mcp__distillate__search_papers with the paper's key topics to find related papers already in the library

3. Cross-reference: If related papers exist, call mcp__distillate__synthesize_across_papers to understand how this paper relates to what's already been read — does it confirm, contradict, or extend prior work?

4. Experiment connections: If experiments are active:

   • Check if the paper's techniques could improve any running experiment
   • Call mcp__distillate__suggest_from_literature for the most relevant experiment
   • If a strong connection exists, call mcp__distillate__link_paper to associate them

5. Extract the tincture: Synthesize a concentrated summary:

   • Core contribution: one sentence on what's new
   • Key technique: the method, described concretely enough to implement
   • Reported results: metrics and baselines from the paper
   • Connections: how it relates to other papers and active experiments
   • Actionable ideas: 1-3 specific things to try based on this paper

6. Offer next steps: promote the paper, steer an experiment with its insights, or add related papers to the library