Free Meeting Recap Generator + How AI Meeting Summaries Work (2026)
Paste your notes or transcript and get a clean meeting recap in seconds, free, in your browser. Plus: what a good recap contains, a copyable template, two real examples, how AI meeting summary tools work (and where they fail), and how to get automatic recaps on Mac without sending audio to the cloud.
A meeting recap is the short written summary you send after a meeting: what was discussed, what was decided, who is doing what by when, and what is still open. A good one takes two minutes to read and saves the next meeting from starting with "wait, what did we agree last time?"
Paste your notes or transcript below to generate a recap, or scroll down for the template, examples and an honest look at how AI meeting summary tools work.
Free meeting recap generator
Paste rough notes or a transcript. Everything runs in your browser — nothing is uploaded.
What a good meeting recap contains
Most recaps fail in one of two directions: so vague they are useless, or so long nobody reads them. The fix is structure. A useful meeting recap has six parts, in this order:
- Header. Meeting title, date, attendees. If someone skims only this line they should know which meeting it was.
- Summary. Two or three sentences on what the meeting was about and where it landed. Not a chronology, a conclusion.
- Decisions. Each decision as one bullet, stated as a fact: "We will launch on 14 October," not "we discussed launch dates."
- Action items with owners and dates. One line per task: what, who, by when. An action item without an owner is a wish; without a date it is a suggestion.
- Open questions and blockers. Anything raised but not resolved, so it lands on the next agenda rather than in someone's memory.
- Next meeting. Date, time, and what needs to be ready for it.
If any of these is missing, the recap is incomplete. If anything else is in it (who said what, the full discussion, the tangents), move it to the transcript and link to it.
Three rules of thumb: write it within an hour while details are fresh, keep it under 250 words, and send it to everyone who attended plus anyone whose work it affects.
Meeting recap template
Copy this into your email, Slack, Notion or Apple Notes and fill in the brackets. It works for internal meetings, client calls and one-on-ones alike.
# Meeting recap: [Meeting title]
Date: [YYYY-MM-DD] · Attendees: [Names] · Recorded by: [Name]
## Summary
[Two or three sentences: what the meeting was about and the outcome.]
## Decisions
- [Decision 1, stated as a fact]
- [Decision 2]
## Action items
- [ ] [Task] — Owner: [Name] — Due: [Date]
- [ ] [Task] — Owner: [Name] — Due: [Date]
## Open questions / blockers
- [Question or blocker, and who will resolve it]
## Next meeting
[Date and time] — Prepare: [what needs to be ready]
A shorter variant for quick syncs: drop the header and summary and send only decisions, action items and open questions. A longer variant for client calls: add a "Context" line above the summary with the account name and the reason for the call.
Meeting recap examples
Two short examples of the template in use. Names and companies are invented.
Example 1: internal product sync
Meeting recap: Q4 roadmap sync Date: 2026-08-20 · Attendees: Sofie, Karim, Lena, Tom · Recorded by: Lena
Summary. We reviewed the three candidate features for Q4 and agreed to ship offline export first, push the calendar redesign to Q1, and run a two-week spike on the sharing feature before committing.
Decisions
- Offline export ships in Q4; calendar redesign moves to Q1 2027.
- Sharing feature gets a two-week spike, decision on 3 September.
Action items
- Karim: write the offline export spec. Due 27 August.
- Sofie: scope the sharing spike and book the review. Due 3 September.
- Tom: update the public roadmap page. Due 22 August.
Open questions
- Does offline export need a paid-tier gate? Lena to check with finance.
Next meeting: 3 September, 10:00. Prepare: spike findings.
Example 2: client call recap
Meeting recap: Acme onboarding call Date: 2026-08-21 · Attendees: Marie (Acme), Jonas (Acme), Pieter (us) · Recorded by: Pieter
Summary. Kick-off for Acme's rollout to the sales team. Acme confirmed a 15 September go-live, wants Dutch and French support from day one, and needs a data-processing summary for their DPO before sign-off.
Decisions
- Go-live target: 15 September, sales team first, support team in October.
- Training will be one 45-minute session per team, remote.
Action items
- Pieter: send the data-processing summary to Marie for the DPO. Due 25 August.
- Jonas: share the list of sales-team users and their languages. Due 28 August.
- Pieter: propose two training slots in the week of 8 September. Due 29 August.
Open questions
- Whether Acme's IT needs an MDM deployment package. Jonas to confirm.
Next meeting: 5 September, 14:00. Prepare: DPO feedback.
Both recaps are under 150 words and take less than a minute to read. That is the target.
How AI meeting summary tools work (and where they fall short)
Every AI meeting summary tool, whether it is a bot notetaker, a browser extension or a native app, does the same three things:
- Transcribe. A speech-to-text model (Whisper, Parakeet, or a vendor's cloud engine) turns the audio into text, ideally with speaker labels and timestamps.
- Summarise. A language model reads the transcript with a prompt or template that says, in effect, "extract the summary, decisions, action items with owners and dates, and open questions." Better tools let you customise that template per meeting type.
- Deliver. The result is pushed to a dashboard, an email, Slack, Notion or Apple Notes.
Done well, this collapses twenty minutes of post-meeting writing into a few seconds and gives every meeting a consistent recap regardless of who was tired or in a hurry. Teams that adopt it report the same thing: fewer follow-up meetings to explain the previous meeting, faster onboarding (read last month's recaps instead of scheduling briefings), and fewer dropped action items.
Where the tools fall short is worth knowing before you trust one:
- Garbage in, garbage out. The summary is only as good as the transcript. A tool that transcribes Dutch at 65% accuracy will produce confident-sounding recaps with wrong names and invented decisions. Check the transcription quality in your language before you judge the summaries.
- Owners and dates are the weak spot. Models are good at "what was discussed" and noticeably worse at "who committed to what by when," especially when the commitment was implicit ("I can look into that"). Always review the action-item list.
- Nuance and internal jargon. The first draft will not know your project code names or that "the Friday thing" means the release. Custom templates and a quick edit fix most of this.
- Too long by default. Many tools produce a page when a paragraph would do. If your recap needs scrolling, tighten the template.
- Where the audio goes. Most AI summary tools upload the audio and the transcript to a cloud service. For client calls, HR conversations or anything under GDPR, that makes the vendor a data processor and you the one responsible for the DPA, the lawful basis and the transfer safeguards.
The practical answer to all five is a human in the loop: generate automatically, skim, fix the owners and dates, send. That is still a fraction of the time it takes to write a recap from scratch, and it is the workflow the tool on this page and MeetMemo are both designed for.
Automatic recaps on Mac without sending audio to the cloud
If you want every meeting to produce a recap like the examples above, automatically, and you are on a Mac, MeetMemo does the whole pipeline on-device:
- Record any meeting your Mac can hear (Zoom, Teams, Google Meet, Slack huddles, phone calls routed through the Mac, in-room conversations) with no bot and no browser extension.
- Transcribe with NVIDIA Parakeet on the Apple Neural Engine, live during the call, with speaker labels. Dutch, English and French are handled out of the box, including Flemish dialect and code-switching.
- Summarise with Apple MLX (Qwen 3 4B by default) on your own Mac into exactly the structure above: summary, decisions, action items with owners, open questions. Custom templates let you shape the recap per meeting type.
- Deliver to Apple Notes in a MeetMemo folder, organised by date and title, plus a local Markdown knowledge base you can search, open in Obsidian, or query through a local MCP server from Claude Desktop or Cursor.
Nothing is uploaded: not the audio, not the transcript, not the summary. There is no account to create, it works offline, and because you are the only party processing the data there is no Data Processing Agreement to sign. Your first three meetings are free with every feature; after that it is €9/month or €79/year. Requirements: macOS 14 or later on Apple Silicon.
Share a recap the way you already share notes: select the note in Apple Notes and send it by Mail or Messages, or copy the Markdown into Slack or an email. For the recording mechanics see how to record a meeting on Mac; for more note structures, meeting notes templates; and for a comparison of which local models write the best summaries, our local LLM meeting summary benchmark.
Frequently asked questions
What is a meeting recap? A short written summary sent after a meeting covering what was discussed, what was decided, who owns which action items by when, and what is still open. Readable in under two minutes.
Meeting recap vs meeting minutes: what is the difference? Minutes are a formal, often chronological record, sometimes legally required. A recap is informal and action-oriented. Minutes document what happened; a recap tells people what to do next.
How long should a meeting recap be? 100 to 250 words, or one summary line plus three short bullet lists. If it needs scrolling, link to the transcript instead.
How do I write a meeting recap quickly? Use the template above, write within an hour of the meeting, or paste your notes or transcript into the generator at the top of this page and edit the result.
What should a meeting recap include? Header (title, date, attendees), summary, decisions, action items with owner and date, open questions, next meeting.
How do AI meeting summary tools work? Speech-to-text turns the audio into a transcript; a language model extracts summary, decisions, action items and open questions into a structured format; the result is delivered to your notes app or inbox. Quality depends on transcript accuracy, the template and a quick human review.
Can I get automated meeting summaries without the cloud? Yes. MeetMemo records, transcribes and summarises on your Mac and saves the recap to Apple Notes. Nothing is uploaded.
Is the recap generator on this page free? Yes. No account, no sign-up. Paste, generate, copy.
Stop writing recaps from scratch
Use the generator above for the meeting you just had. For every meeting after that, let your Mac do it: download MeetMemo, press record, and the recap is in Apple Notes before you have closed the call window.