CAPTUREASK · PRACTICAL GUIDE
Area capture: include the evidence and leave out distractions
Updated September 13, 2026 · Based on the downloadable v0.3.6 package
A useful crop keeps the detail you want explained together with its labels and context. Cropping too tightly can make a screenshot less useful, even when it looks cleaner. Here is a repeatable workflow for CaptureAsk v0.3.6.
Drag directly over a page
- Open a normal web page and scroll the relevant content into view.
- Open CaptureAsk and select the AI service you want to use.
- Choose Drag to capture & send. The popup closes and a selection overlay appears on the page.
- Press and hold the pointer at one corner of the desired rectangle, drag to the opposite corner, and release.
- The extension crops the visible tab and opens the selected AI. Check that the resulting attachment contains the intended detail.
Press Escape to cancel while the selection overlay is active. A selection smaller than roughly five pixels in either dimension is ignored by this version. The mode selects part of the visible page; it does not scroll and stitch a long document.
Example: crop a chart without losing its meaning
The following invented figures illustrate why labels matter; they are not real business results.
| Week | Opened | Resolved |
|---|---|---|
| Week 1 | 120 | 108 |
| Week 2 | 135 | 132 |
A crop of just “120, 108, 135, 132” loses the distinction between opened and resolved tickets. Include the title, column labels, units and period. If the page contains a legend or active filter, include it too. Exclude unrelated sidebars and personal details.
Using only this table, compare opened and resolved tickets for each week. Show the arithmetic. Do not infer the total backlog because the starting backlog is not provided.
A checkable answer gives a net increase of 12 tickets in Week 1 and 3 in Week 2. It should not claim the final backlog is 15 unless the initial backlog was zero. This is the kind of boundary you can state explicitly in your prompt.
Example: an error message
Keep the complete error text, the application name and the operation that failed. Remove tokens, account identifiers and private paths. Ask the AI to separate what the message proves from possible causes. A screenshot of “failed” alone rarely supports a useful diagnosis.
When dragging does not work
Chrome restricts extensions from injecting scripts into some pages, including internal browser pages and the Chrome Web Store. Try the same action on an ordinary HTTPS page. If the overlay appears but you get no image, reopen CaptureAsk: v0.3.6 stores a short-lived area preview in the browser session. If available, inspect and save it before retrying.
For a preview-first alternative, choose Current tab, then use the preview’s crop action and apply the rectangle. This is a different workflow from dragging directly over the web page. If the page changes while selecting, stop its animation or wait for it to settle and capture again.
Before sharing
Check the final image, not just the selection rectangle. Leave enough context to understand the question, but remove names, private conversations and unrelated records. If automatic attachment fails, save the preview and use the provider’s file picker. Copy-and-paste can be another fallback when clipboard access is available.
Implementation details were checked against the downloadable extension code. AI provider interfaces and account limits can change; automatic attachment is not guaranteed. Examples on this page are illustrative.