Do I have to save the screenshot before attaching it?
No. The annotated image and your transcribed words go onto the clipboard together; one paste delivers both. The file is also kept in a local folder in case you need it again.
use case · bug reports
Everyone hits bugs; nobody enjoys writing them up. Crop the broken thing, say what happened while you circle it, and paste. The screenshot shows what broke; your words tell how — and they arrive together, in the tracker, the team chat or an email.
Drag over the broken part — the error message, the misaligned panel, the wrong total. Recording starts as you drag.
Arrow the button you pressed, box the message that came back — and say what you did and what you expected instead. The steps get spoken, not typed.
Jira, GitHub, Linear, a team chat, an email to the vendor — the annotated screenshot and your transcribed report land together, in one paste.
Nothing to transcribe afterwards: what you said becomes ordinary editable text — read it over, add the version number, fix a word, then submit. If it came out rambly, the Polish voice mode tidies the grammar and drops the filler before it lands — on the built-in local model or your own Ollama server, or through your own OpenAI or Claude account with Pro.
QA files dozens a day, but everyone runs into things that are broken: the form that won't submit, the invoice that came out wrong, the page that renders scrambled. It's the same gesture whoever you are — crop it, say it, send it to whoever can fix it. And when a customer reports it first, support teams pass it along the same way.
Paste the same report into a terminal-based tool and delivery adapts: the terminal gets your words plus the screenshot's file path, which a tool like Claude Code can open. With Pro, a connected coding agent can go one step further and pull your latest capture itself, or ask you for a fresh one. Connect your AI apps →
Bug screens are full of live data: accounts, orders, names. By default everything stays on your machine — your speech becomes text on your device, screenshots stay in a local folder, audio isn't uploaded to anyone, and the app tracks nothing about you. What reaches the tracker is exactly what you paste. The full network-use table →
No. The annotated image and your transcribed words go onto the clipboard together; one paste delivers both. The file is also kept in a local folder in case you need it again.
Yes — Alt+1 captures and annotates without recording. Type the steps next to the image, or leave the picture to speak for itself.
The app you pasted into could not take an image, so it got your words plus the file's location. Open History from the button and click the entry to deliver it again, or set a fixed preference in Settings.
No — only the region you dragged over, with your markup. Nothing outside that crop leaves your computer.
Free for QA, and for everyone else who finds the bugs. Related: voice for support teams, asking AI about your screen, design & document feedback and all use cases.