Published
September 3, 2026
in

AI that writes your weekly status update from your meetings, calendar, email, tasks, and screen

Carson Eisner, Growth Engineer
Carson Eisner
Growth Engineer

Littlebird is an AI that writes your weekly status update from the week it already saw: the meetings it transcribed, your calendar, your email, your task list, the Notion pages you work in, and whatever was on your screen while you worked. You save a five-heading prompt once as a weekly Routine, and the draft arrives Friday at 4:00. The prompt below asks for a source on every line, so a two-minute check is enough before you send. The free Basic plan includes a limited number of active Routines. Below: what each source contributes to each heading, the exact prompt, a sample draft, what it can't see, and how it differs from generators and tracker agents.

What does the AI write the update from?

It's Friday at 4:45 and the update is due at 5:00. The week is in six places: a PR queue, a ticket board, two threads, a standup, and a calendar full of calls. An AI can only write the update from what it can reach, so the source list is the whole decision. Here is what Littlebird reads, and which heading each source feeds.

  • Your screen. What Littlebird reads: The text and elements of the active window, every few seconds: the PR you reviewed, the doc you edited, the ticket you had open, the thread you answered. Heading it feeds: Done, Next.
  • Meeting Notes. What Littlebird reads: Every transcribed meeting's Decisions, Action items split into yours and everyone else's, and Risks and open questions. No bot joins the call. Heading it feeds: Decisions, Blocked, Risks.
  • Google Calendar. What Littlebird reads: What's on your calendar, this week and next. Heading it feeds: Done, Next.
  • Gmail. What Littlebird reads: Open loops: the thread you're waiting on, the reply you owe. Heading it feeds: Blocked.
  • Todoist or TickTick. What Littlebird reads: Completed tasks this week, and tasks by due date for next week. Heading it feeds: Done, Next.
  • Notion. What Littlebird reads: Any page or database view you point it at: search, pull a page's full content, query a view. Heading it feeds: Done, Next.
  • Outlook. What Littlebird reads: Your Microsoft calendar and mail, searched by keyword, sender, or date. Heading it feeds: Done, Blocked.

The screen row is the one no other tool has. Littlebird runs on your computer and reads the text and elements of your active window via macOS accessibility permissions, every few seconds, building a private, encrypted index of recent activity. It's not a screen recorder: no video, no screenshots, no keystroke logging. The PR you spent Tuesday on is in the draft because you had it open, with nothing to log. The other rows are optional connections. Gmail, Google Calendar, Notion, Todoist, and TickTick are verified in the app; Outlook is documented on Littlebird's support site. Each adds what the screen alone would miss: the task you ticked off from your phone, the email that landed while you were on a call.

Do you have to connect every app?

No. It works automatically by paying attention to your screen. Integrations are optional, for deeper access. Connect the ones that hold work the screen wouldn't catch, and skip the rest. If your tickets live in Jira, Linear, Asana, or ClickUp, the route we can vouch for here is the screen: the tickets you opened this week were read as you worked them. Tickets you never opened stay out of the draft. That's the honest trade against a tracker-connected agent, and it's covered below.

What prompt does the Routine run?

A Routine is a saved prompt that runs on a schedule, daily, weekly, or monthly at a chosen time, and pushes the result back to you. This is the one to save. It names the sources, so the draft pulls from each of them instead of from whatever was loudest.

Write my weekly status update for the week of [dates]. Draw on my meeting notes from this week, my calendar for this week and next, my email, my completed and upcoming tasks, the Notion pages and databases I work in, and the documents, tickets, and pull requests I had open. Use exactly five headings, in this order: Done, Next, Blocked, Decisions, Risks. Under Done, list three to five things that shipped or closed. Under Next, list what I'm doing next week in priority order, including meetings already booked. Under Blocked, name who I'm waiting on and since when. Under Decisions, list anything settled this week that other people should know. Under Risks, say what might slip and why. After every line, name the source in parentheses: the meeting, the thread, the task, or the document. Keep it to one screen, plain sentences, no summary paragraph. If something is unclear, leave a bracketed question for me instead of guessing.

What does the draft look like?

One screen, with a source after every line. Placeholders where your week goes:

  • Done. Shipped [feature] to production (pull request, open on screen Tuesday). Closed [ticket] after the [customer] call (Meeting Notes, Wednesday, your action items). Finished [task] and [task] (completed tasks, Todoist).
  • Next. [Task] first, due [day] (tasks by due date). [Meeting] on Tuesday to settle [question] (calendar, next week).
  • Blocked. Waiting on [name] for [thing] since Wednesday (email thread, no reply; raised again in Thursday's standup).
  • Decisions. We're doing [decision] instead of [alternative] (Meeting Notes, [meeting], [day], Decisions).
  • Risks. [Deliverable] may slip a week if [dependency] lands late (Meeting Notes, Risks and open questions; the dependency's ticket, open on screen Thursday).

The sources are the point. A manager who reads a dozen of these trusts the one that can be checked, and you can check this one in the time it takes to read it. A line with no source is the line to look at before you send.

How do you set it up?

  1. Let the week happen. Nothing to clip, nothing to log. Meeting Notes captures calls from an automatic pre-meeting prompt, or auto-starts for supported meeting apps if you turn that on in Settings > Meetings. If you already have transcripts from another notetaker, import them in Settings > Import and they become searchable Meeting Notes too.
  2. Connect what the screen won't catch. Settings > Integrations: Gmail, Google Calendar, Notion, Todoist or TickTick, Outlook. Skip any you don't use. The Routine works from the screen and your meetings alone.
  3. Save the Routine. Paste the prompt, choose weekly, pick Friday at 4:00, and choose push, email, or both for when the draft is ready. Littlebird ships ready-made starting points: a project status compiler for one person's week and a team status compiler for a manager rolling several updates into one.
  4. Review, then send it where status goes. Two minutes: which risk actually worries you, which blocker needs escalating, what to leave out. Then ask Chat to put it where your team reads it. Littlebird can create a new Notion page from a chat message, and through Gmail it can send email on your behalf. Past updates stay searchable in Chat: what did my update say two Fridays ago, what did I promise Dana.

One setting to know: Routines auto-pause by default if you stop opening their output, so the update won't keep generating into a void. Open it on Fridays and it keeps running; resume anytime from the Routines dashboard.

What can't it see?

Say this part out loud, because an update built on a gap is worse than one you wrote by hand. Littlebird reads the active window on the computer it's installed on, so work done on a second machine isn't in the draft, and neither is anything done on your phone: the iOS and Android apps let you ask about the context your desktop collected, and they don't capture activity from the phone. Anything you excluded stays out: specific apps, specific website domains, or whole categories such as Banking or Health, plus any stretch where you paused capture for 5 to 60 minutes or until the next launch. Tickets you never opened and threads you never read aren't in the draft either; they were never on your screen. And it doesn't decide what counts as a risk. That's the judgment you add in the two-minute review.

Where what it saw ends up: the app runs on your computer, and the memory it builds is encrypted and stored in the AWS cloud (US East), AES-256 at rest and in transit, SOC 2 Type II compliant, and never sold. You can delete the last hour, the last day, or everything. Dedicated deployments for larger rollouts are arranged with our team as part of an enterprise agreement. The full picture is in Is Littlebird safe?

How do generators and tracker agents differ?

By source, which is the only thing that differs.

  • Free status report generator (a web form). Writes from: The notes you type into the form. What you still do: Collect the week first, then reformat the output.
  • ChatGPT or Claude. Writes from: What you paste in. What you still do: Gather PRs, tickets, notes, and last week's update; paste; repeat every Friday.
  • Notion's weekly status reporter agent. Writes from: Notion pages and Slack conversations, on a schedule (notion.com, read October 6, 2026). What you still do: Keep the work in Notion and Slack; add what happened on calls and in email.
  • Littlebird's weekly Routine. Writes from: The active window, Meeting Notes, Google Calendar, Gmail, Todoist or TickTick, Notion, Outlook. What you still do: Read the draft, add judgment, send.

ChatGPT is a powerful general assistant with no context about your work; you supply it each time by copy and paste, and pasting internal threads into a general assistant is a question your security team may have views on. Notion's agent is a real tracker-side answer: it runs on a schedule, reads Notion pages and Slack conversations for this week's updates and next week's plans, creates a new report page with This week, Next week, and a short TLDR, and links back to sources so the team can verify details (notion.com, read October 6, 2026). It writes well from work that lives in Notion and Slack, and nothing from the call where the decision actually moved. Littlebird's difference is that the record was kept for you, across the apps, with the meeting in it.

What does it cost?

The free Basic plan, $0, includes limited active routines, limited meeting notes, and connecting calendar and email, so the weekly Routine runs on the free plan with the screen and your meetings behind it. Plus is $17 per month billed annually or $20 per month billed monthly, with a 14-day free trial, and adds more routines and unlimited meeting notes, which matters once the standups outrun the Basic cap. Those are the figures from littlebird.ai/pricing as recorded in our fact base; the live pricing page is the final word, and the full ladder is in our pricing guide.

Who gets the most out of it?

Engineers first. A senior engineer's week is spread across a PR queue, a ticket tracker, a standup, and threads, and the archaeology of pulling it back together is exactly what an assistant that watched the week removes; more in the engineers guide. Consultants with several clients get a per-client version of the same draft. Managers get the roll-up, with the team status compiler as the starting point. Founders get the longer cousin, the investor update, and anyone who wants the daily version has the daily work journal; a weekly update is five of those stitched together. In a Littlebird user survey (2026), 43% of subscribers estimated saving three or more hours per week, and Friday's update is one of the easiest of those hours to find.

Common questions

Can AI write my weekly status update automatically?

Yes. Littlebird drafts it on a weekly Routine from the meetings it transcribed, your calendar, your email, your completed tasks, your Notion pages, and what was on your screen, in the five-heading template you set once. A general assistant writes it from what you paste. A tracker agent writes it from the tracker.

Do I need to connect all my apps?

No. It works automatically by paying attention to your screen. Integrations are optional, for deeper access. Connect Gmail, Google Calendar, Notion, Todoist, TickTick, or Outlook to add what the screen wouldn't catch.

Can it pull from Jira, Asana, Linear, or ClickUp?

The route we can vouch for here is the screen. Littlebird reads the active window, so the ticket you had open is in the draft, and so is the thread the ticket never mentioned. Tickets you never opened aren't, and a tracker-connected agent covers those.

What's the best template for a weekly status update?

Five headings, in this order: Done, Next, Blocked, Decisions, Risks. One screen, same shape every week, and a source after every line so the reader can check it.

Will it send the update for me?

The Routine delivers the draft to you by push or email. Sending is your call: ask Chat to turn it into a Notion page, or, with Gmail connected, to send it on your behalf.

Will it keep writing updates nobody reads?

No. Routines auto-pause by default when you stop opening their output. Resume anytime from the Routines dashboard.

Friday's update, from the week you already had.

Save the Routine once. The draft lands at 4:00 with its sources. Free Basic plan, no setup.

Get Littlebird

Product facts come from our maintained internal fact base and were re-checked October 6, 2026; the connections named on this page are limited to the ones that fact base records as verified in the app or in Littlebird's support docs. Notion's weekly status reporter agent is described from notion.com, read October 6, 2026. Survey figure: Littlebird user survey, 2026.