The screens a PMcan't copy from
A product manager's day is full of text trapped in pixels: pasted screenshots, phone photos, read-only dashboards, locked PDFs. Khint's Capture pillar gets the words out (extract them to your clipboard, or question the image with vision chat), then hands them to an Action so they end up in a ticket, not retyped by hand.
Retyping a screenshot
is the small tax that adds up
Nobody schedules “copy that number off the screenshot”, but it happens a dozen times a day: a figure from a dashboard you can't click into, an error string in a crash dialog, a requirement someone photographed off a whiteboard. Each one is a tiny detour: squint, retype, hope you got it right. Capture is one of Khint's three pillars (Agents, Capture, and Memory), all behind one palette, so reading a screen and acting on it is the same gesture.
From pixels to a paste-ready ticket
Five ways the Capture pillar shows up in a PO or BA's day.
Decide how you want to read the screen
Open the palette with Cmd+Shift+K. Capture lives in two rows: Extract text when you just need the words on your clipboard, and Capture & ask when you want to question the image across a few turns. No separate hotkey to remember: both sit in the same palette as your Actions. How Capture works →
Pull text out of anything on screen
Pick Extract text, drag a box around the region (a Slack screenshot, a stakeholder's emailed mockup, a number buried in a read-only dashboard) and the OCR'd plain text lands on your clipboard. A small overlay confirms the grab with a one-line preview, so you stay in the doc you were writing.
Drop a file when it isn’t on screen
Not everything is on a monitor. A phone photo of a whiteboard, a scanned page, an exported chart: drag it into the Capture tab's drop zone or pick it from disk. PNG, JPG, WebP, HEIC and more all come back as clean text, the same as a region capture. Turn a whiteboard photo into requirements →
Ask the image, don’t just copy it
When a screen is dense (a competitor's flow, a metrics panel, a UI you're reverse-engineering) use Capture & ask. The screenshot pins in the chat window and stays in context, so you can ask “what fields are required here?” then “rewrite that as acceptance criteria” without starting over. Extract requirements from a screenshot →
Pipe the text into a ticket, not a dead end
Captured text uses the same paste path as a Khint Action, so it feeds straight into one: shape it into a user story, summarize it for a ticket, or chain a couple of steps. With Atlassian connected on the integrations page, you can push the result to Jira or Confluence from the same shortcut. Screenshot to user story →
Extract text or Capture & ask: which one?
Extract textis the fast grab: the OCR'd words land on your clipboard and you paste them wherever you were working. Reach for it when the screen is simple and you just need the characters. Capture & askis the read-and-reason path: the image pins in a chat and stays in context, so you can interrogate a dense screen (what's required, what a chart implies, how to phrase it as criteria) across several turns. Both start the same way: open the palette, drag a box. From there, the captured text behaves like any other input, so it can flow into a saved Action or a multi-step workflow, and out to a Jira bug ticket if that's where it belongs.
Common questions
What does a product manager actually use OCR for?
All the text a PO or BA gets handed that they can't select: a stakeholder's screenshot pasted in Slack, a phone photo of a wireframe, an exported chart in a slide, a non-selectable PDF spec, a legacy screen in a live demo. Khint's Capture pillar gets that text out: open the palette with Cmd+Shift+K, pick Extract text, drag a box, and the plain text lands on your clipboard, ready to paste into a ticket or a planning doc. Capture is one of Khint's three pillars alongside Agents and Memory.
Does it work on files, not just what is on my screen?
Yes. Besides region capture, the Capture tab has a drop zone: hand it a PNG, JPG, WebP, HEIC, GIF, BMP, or TIFF, or pick one from disk, and the text comes back the same way. So a phone screenshot a teammate AirDropped you, a scanned page, or an exported dashboard image all become editable text without retyping.
Can I ask questions about a screenshot instead of just copying the text?
That is Capture & ask. Pick it from the palette, drag a box, and the image lands pinned in a chat window. It stays in context across turns, so you can ask follow-ups without re-uploading: explain a chart, summarize what a busy screen is showing, or translate a label. It is the read-and-reason path, where Extract text is the one-shot grab.
Where do my captures go? Is this private?
A capture is sent to Khint's vision service to read the text, then the result lands on your clipboard and in your local history. History bodies (the input and output text) live in a local SQLite database on your Mac and are never synced to Khint's servers. Capture also has its own daily allowance separate from your AI Action quota, so reading screens doesn't eat into the actions you save for writing tickets.
Stop retyping screenshots
Free with 10 AI actions and 5 captures per day. No credit card. Capture lives in the palette, one shortcut away.