Extract Action Items from Meeting Notes
Paste meeting notes or daily bullets and get a structured checklist with owners plus a decision log. Deterministic, private, and no LLM invention.
Turn messy meeting notes into a checklist
Paste meeting notes, standup notes, or daily bullets and this tool extracts the lines that look like action items plus the lines that record decisions. It runs entirely in your browser and uses deterministic rules: explicit ACTION: / task: markers, @handles, owner assignments like Alice will send the deck, trailing owner tags, imperative task verbs, and decision words such as Decided, agreed, or resolved.
Because it is not an LLM, it does not invent unstated tasks or rewrite your meeting into a summary. The output is a structured task list you can copy into a project tracker.
Choose Markdown for a checklist, or JSON when you want to feed the result to a script. Keep the default Group by type to show one action checklist with owners inline, or choose Group by owner to make headings for each person. Turn off Include decisions when you only need tasks.
Worked example
Input:
ACTION: update the roadmap
Alice will send the deck
Book the venue @bob
Decided: ship on Friday
Output:
## Action Items
- [ ] update the roadmap — _Unassigned_
- [ ] send the deck — **@Alice**
- [ ] Book the venue — **@Bob**
## Decisions
- ship on Friday
FAQ
Does this use AI to infer tasks from vague prose?
No. It is intentionally deterministic: it only extracts lines with explicit action or decision signals. That keeps it predictable and avoids hallucinated tasks, but it also means subtle commitments hidden in prose may be skipped.
How are owners detected?
Owners come from @handles, leading assignments such as Alice will send the deck or Bob to book the room, owner markers such as owner: Sam, and short trailing tags such as Prepare slides (Carol). Pronouns like we are not treated as owners.
What counts as a decision?
Lines with decision markers such as Decided:, decision, agreed, resolved, approved, consensus, or final call go into the Decisions section. Prefixes such as Decided: are stripped so the decision reads cleanly.
Why did a line not appear in the output?
It probably did not have an explicit signal. Add ACTION:, task:, an @owner, a leading owner assignment, or a listed action verb such as send, review, schedule, update, finish, fix, or verify.
Limits & edge cases
- Input is read one line at a time. Put each bullet, action, or decision on its own line.
- The extractor cleans bullets, numbering, and checkbox prefixes, but it does not parse nested agendas or tables.
- It does not extract due dates, priorities, transcript speakers, or implicit commitments. Those require semantic/LLM handling or a separate purpose-built parser.
- If no action item or decision is found, the tool reports a clear error instead of returning an empty list.
Developer & Automation Access
Run it from the terminal
Same engine as this page, headless — via the gizza CLI:
gizza tool action-item-extractor "ACTION: update the roadmap
Alice will send the deck
Book the venue @bob
Decided: ship on Friday"New to the CLI? Get gizza →
Open it by URL
Pre-fill and auto-run this tool with query parameters — the names match the API/CLI:
https://gizza.ai/tools/action-item-extractor/?input=ACTION%3A%20update%20the%20roadmap%0AAlice%20will%20send%20the%20deck%0ABook%20the%20venue%20%40bob%0ADecided%3A%20ship%20on%20Friday&format=markdown&group_by=type&include_decisions=trueMachine-readable descriptor: tool.json — title + parameters JSON Schema for agents.