Loading page
About this tool
Scan raw meeting notes for decisions, actions, owners and due dates, then build a prompt that fills the gaps it found.
Meeting Notes Prompt Builder classifies every line of raw notes as a decision, a commitment, an open question or context, then checks whether each commitment names an owner and carries a due date. It reports the gaps as counts — how many commitments have nobody attached, how many have no date, how many questions were left open — and writes those exact figures into a prompt that instructs the assistant to mark them rather than fill them in. Analysis happens in the browser, so nothing is uploaded.
Open Meeting Notes Prompt Builder on AltFTool — it loads instantly in your browser.
Add your input to the workspace.
Adjust the options until the result looks right.
Copy or download the output and put it to work.
Missing owners, missing dates and unresolved questions are reported as numbers.
The prompt explicitly forbids assigning an owner or a date that the notes do not contain.
Classification is plain pattern matching in your browser — notes never leave the page.
Three things: decisions, commitments and open questions. A decision states what was settled, a commitment names one owner, one deliverable and one date, and an open question names who can close it. Discussion that produces none of the three is context and can usually be compressed to a line.
Because summarisation models fill structural gaps by default — asked for an actions table, they will produce one even when the notes contain no owner or date. The fix is to state the gaps explicitly in the prompt and require them to be marked as unknown, which is what this tool writes for you.
About a minute per 238 words for adult non-fiction read silently, which is the mean from Brysbaert's 2019 meta-analysis of 190 reading studies. That makes a 500-word set of notes roughly two minutes, which is worth knowing before you paste them into a channel where nobody will read them.
No. The line classification, owner detection and date detection all run as pattern matching in your own browser, and the generated prompt is only copied to your clipboard when you press the button. What you subsequently paste into an AI assistant is governed by that assistant's own privacy terms.