Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.runwita.com/llms.txt

Use this file to discover all available pages before exploring further.

This is the loop Runwita exists for: get a meeting in, let the AI read it, review what came out, save it onto the right journey. The whole thing takes about a minute once you have a transcript.

​
What counts as “a meeting”

Anything text-based that captures what was discussed. Runwita accepts:
  • A transcript from MeetingScribe (or any other recorder that exports text segments).
  • Pasted notes from a meeting you took by hand.
  • An exported .eml email thread.
  • A PDF or Word doc with meeting notes.
  • An image of a whiteboard or a screenshot, processed via OCR on macOS.
For your first run, the cleanest path is a transcript or pasted notes. Email and image flows work the same way but layer extra parsing on top.

​
The Inbox

Everything starts in the Inbox.
TODO: screenshot of Inbox with one ready item
The Inbox shows new captures at the top, ready-for-review items below. New items are anything that hasn’t been processed yet, the recorder dropped a transcript in, you pasted some notes, you imported a .eml file. Ready items are ones the AI has read and you haven’t filed yet. If you’ve connected MeetingScribe (or any folder Runwita is watching), transcripts appear here automatically. Otherwise, paste or drop a file:
  • Drag and drop a .txt, .md, .pdf, .docx, .eml, or image file onto the Inbox.
  • Paste text directly using the Capture box at the top.
  • Quick capture writes to the scratchpad, a single rolling note you can flush into the Inbox when you’re ready.

​
Process with AI

Click an inbox item to open it. You see the transcript on the left and a column on the right marked Runwita reading.
TODO: screenshot of the Process with AI screen, with date editor visible
Two things to set before clicking Process with AI:
  • Meeting date. Defaults to the file’s modification date. The AI uses this to resolve relative phrases like “by Friday” in extracted action items. If the date is wrong, the due dates on every action will be wrong. Pick the right date here.
  • That’s it. Just the date.
Click Process with AI. Runwita streams the extraction in real time, you’ll see sections, decisions, and actions populate as they come back. Most meetings extract in 10 to 30 seconds. Long merged extractions (multiple transcripts stitched together) can take up to a minute. When extraction finishes, the panel shows a green ✓ Complete badge and a summary card. Two buttons sit at the bottom:
  • ↺ Reprocess with AI, throws away this extraction and drops you back to the start. Use this if the date was wrong, or you want to try a different model.
  • Review & Save →, takes you to the review screen.

​
Review and Save

The review screen is your last chance to edit before the engagement is created.
TODO: screenshot of Review & Save with journey picker, date editor, and extracted brief
Three things to confirm:
1

Journey

Runwita auto-suggests a journey based on the meeting content. You’ll see one of:
  • High confidence, the AI is sure. Already selected.
  • Medium confidence, plausible match. Already selected, but worth a glance.
  • Low / no match, you’ll need to pick or create one.
Click Change to pick a different existing journey, or + New to create a fresh one inline.
2

Engagement date

The AI’s best guess at the meeting’s date, prefilled. Override here if it’s wrong. Whatever’s in this field at save time becomes the engagement’s date on the journey timeline.
3

The extracted brief

Title, summary, sections, decisions, actions, attendees. Everything is editable inline. Hover any line to see which part of the transcript it came from in the right pane.
When everything looks right, click Save to journey →. Runwita does three things in one go:
  1. Creates the engagement on the journey.
  2. Runs topic matching, deciding whether each section maps to an existing topic on the journey or creates a new one.
  3. Pushes any TickTick-flagged actions to TickTick (if connected).
You’re back at the journey, with the new engagement at the top.
TODO: screenshot of journey page with new engagement at top

​
What just happened

In the database, Runwita just did all of this:
  • Created an engagement row linked to the journey.
  • Stored sections, decisions, actions, attendees as related rows.
  • For each AI-detected topic, either created a new topic or attached a touchpoint to an existing one.
  • If it was a transcribed meeting, marked the inbox item as processed.
The journey page now shows this engagement in its history, and the topics list (left rail on the journey) reflects whatever was discussed.

​
What’s next

Concepts: journeys

The deeper logic behind journeys, why they’re shaped the way they are.

Concepts: topics

How threads of discussion span meetings.