Skip to content
workflowintermediateFree

Personal knowledge synthesis

Turn a pile of highlights and notes into connected ideas, tensions worth thinking about and atomic notes for your second brain.

You'll need: Notes or highlights exported as text (Kindle, Readwise, Notion, Obsidian, etc.)

Fill in your details

Values are inserted into every prompt below. Nothing you type leaves your browser.

3 field(s) left — unfilled variables stay as {{placeholders}}.

  1. 1Cluster the notes

    Paste up to about 50 notes. Number them if they are not already numbered so the model can refer back.

    Here are my notes and highlights on {{theme}}. Each is numbered.
    
    {{notes}}
    
    Group them into 4-8 clusters of related ideas. For each cluster: a name written as a claim, the note numbers in it, and a two-sentence summary in plain language. List any notes that don't fit anywhere separately; don't force them into clusters.
  2. 2Find tensions and connections

    This is where synthesis happens. Disagreements between your sources are usually the most useful thing to think about.

    Across the clusters, identify:
    1. Tensions: pairs of notes or clusters that disagree or pull in different directions. Quote both sides.
    2. Surprising connections: ideas from different clusters that combine into something neither says alone.
    3. Challenges to my current view, which is: {{current_view}}. Which notes most strongly challenge it?
    4. Three questions I should investigate next.
  3. 3Write atomic notes

    Paste the output into your notes app. Add your own sentence to each note before saving; that's what makes it yours.

    Write one atomic note for each cluster and each tension, in Markdown, using this format:
    
    # [Title as a claim]
    [3-5 sentences in my words, no jargon]
    Sources: [note numbers]
    Related: [[other note titles]]
    Open question: [one question]

Example input

Theme: habit formation. 38 Kindle highlights from three books and two articles. Current view: environment design beats willpower.

Example output

Tension: "Identity first" vs "Action first" Note 7: "Every action is a vote for the type of person you wish to become." Note 22: "Change behaviour first; beliefs follow." # Small actions build identity, not the other way round Several sources agree that a new self-image follows repeated behaviour more reliably than it precedes it... Sources: 7, 15, 22 Related: [[Environment beats willpower for starting]] Open question: Does this hold for habits people stop doing, not just ones they start?

Troubleshooting

  • Clusters are too broad: ask for more, smaller clusters or set a minimum of five.
  • It summarises instead of connecting: insist on direct quotes for each tension.

Known limitations

  • The model can only connect what you paste; it does not know your earlier notes unless you include them.
  • Synthesised claims can overstate what sources say; keep the source numbers and check before quoting.

More from ChatGPT / Claude / Gemini