Photograph the board,walk out with requirements
A requirements workshop ends with a whiteboard full of sticky notes, boxes, and arrows, and a photo on your phone that nobody ever turns into a document. Khint reads that photo with a vision chat you can keep questioning, sorts it into decisions, requirements, and open questions, and files the write-up in Confluence or Jira, all from one keyboard shortcut.
The workshop was great,
the photo is where it dies
Everyone leaned in, the board filled up, the decisions felt clear. Then the room empties and all that thinking is a slightly blurry photo in your camera roll, and the write-up that turns it into requirements is the chore that slips a week, until half the arrows no longer mean anything. Khint closes that gap: Capture reads the photo and lets you interrogate it, an Agent you tune once sorts it into a workshop record, Memory keeps the whole session in context, and when the draft is ready the integrations file it without leaving the page.
A board, turned into a record
One photographed whiteboard, before and after.
- yellow: guest checkout MUST (no signup)
- pink: save cart for later?
- → address validation (blocks pay)
- box: 1-click reorder (phase 2)
- ?? tax service rate limit ??
- Decision: guest checkout ships without an account.
- Requirement: the system shall validate the address before payment.
- Requirement: the user shall reorder a past order in one click (phase 2).
- Open question: can we save a cart for later?
- Open question: tax service rate limit, confirm with backend.
Illustrative: your draft follows whatever structure you wrote into the Action prompt, and you stay the judge of what the board actually said.
One photo, start to a write-up
Closing out an elicitation workshop, step by step.
Open a session for the workshop
Before you start writing up, open a Memory session: call it “Onboarding redesign workshop”. Every capture and Action you run is logged to it, and Khint keeps a compact running summary. So if the board spilled across three photos, the model keeps the whole workshop in context as you work through them. How Memory works →
Drop the photo into Capture
AirDrop the whiteboard photo from your phone, then drop it into the Capture tab, a HEIC straight off the camera roll works the same as any screenshot. Pick Capture & ask and the image lands pinned in a chat window, so you can question it without re-uploading between turns.
Read the board with the vision chat
Ask in plain language: “Transcribe every sticky note and boxed item on this board, grouped by colour.” Then keep going: “What does the arrow from ‘guest checkout’ point to?”, “Which notes look like open questions versus decisions?” The photo stays in context, so you interrogate and correct it in one thread until the board is fully transcribed.
Shape the transcript into requirements
Hit Extract text for the clean transcript on your clipboard, then select it and run a saved Agent: “sort this workshop board into Decisions, Functional requirements, and Open questions, each as a bulleted list.” Describe the Action once and let the editor Generate the prompt: every workshop after gets the same structure. Clone the Backlog Refiner starter pack from the Agents page if you want a working requirements Action to start from.
File the write-up where the team will read it
With Atlassian connected, the palette can publish the result: Confluence Create page posts the full workshop record (decisions, requirements, parking lot) as one page, while Jira Create issue files the action items as tickets, one per run so each is deliberate. The page or issue key comes back to paste into the thread, and the whole workshop stays in your Memory session as a record of what the board said. How integrations work →
Reading the board is just the first step
The structured write-up is reusable text: paste it into a ticket, feed a single requirement into a follow-up Action that writes acceptance criteria, or chain it. Build a workflow of up to five steps in series: an Agent that sorts the board, a second that drafts the requirements, then a create-page or create-issue step at the end. And because Khint ships an MCP server, Claude Desktop or Codex can read the workshop session directly, the photo you captured, the requirements you drafted, the tickets you filed, without copy-paste.
Common questions
How do I get a whiteboard photo into Khint?
Take the photo with your phone, then AirDrop or save it to your Mac and drop the file into Khint's Capture tab. PNG, JPG, WebP, HEIC, GIF, BMP, and TIFF all go through the same flow as a live screen capture, so the photo straight off your camera roll works without converting anything. You can either Extract text for plain text on your clipboard, or pick Capture & ask to open a vision chat where the image stays pinned and you can interrogate it across turns.
Can it actually read messy workshop handwriting and arrows?
A whiteboard photo is rarely clean: handwriting, sticky notes, boxes joined by arrows. Capture & ask pins the image in a vision chat so you read it conversationally instead of trusting one pass: ask it to transcribe each cluster, ask what an arrow seems to connect, and correct it in the same thread when it misreads a scrawl. You stay the judge of what the board actually said; Khint does the transcribing and structuring so you are not retyping sticky notes one by one.
How is this different from extracting requirements from a screenshot?
A screenshot is a finished digital screen (a mockup or a legacy app) where the fields are already laid out. A workshop whiteboard is the opposite: raw, collaborative thinking captured live, where the output is the decisions made, the requirements agreed, and the questions left open, not a screen inventory. This flow is tuned for that write-up, turning the board into a workshop record, rather than reading a UI. If you were handed a screen instead, see extract requirements from screenshots.
Where does the photo and my workshop text go?
The image goes to the vision model so it can read the board; after that, Actions send only the text you selected, plus, if a Memory session is active, the compacted session summary. Nothing else leaves your device silently, and your Jira or Confluence credentials sit in your OS keychain. Khint's free tier covers 10 AI actions and 5 captures per day with no credit card; paid plans start at €7/mo, with Pro at €29/mo for 100 AI actions a day; every plan includes all the work integrations.
Try it on your next workshop photo
Free with 10 AI actions and 5 captures per day. No credit card. Drop the photo, read the board, keep the requirements.