Cutline

Screenshots named from the page they came from — the product, ticket, or error — instead of a timestamp.

Every screenshot you take is called Screenshot, then a timestamp. Three weeks later you need the one with the error in it, and you are opening files at random. Cutline names the file from what is actually on the page. amazon-lenovo-legion-5i-gen-9-79999.png jira-Q-413488-payment-gateway-timeout.png mendix-bouncycastle-invalid-cipher-text-error.png The filename is the only piece of context that survives contact with the real world. It still works after the file is dragged into Slack, attached to a ticket, or emailed — no app, no database, no index required. CAPTURE • Region — drag over a frozen copy of the page, so nothing shifts under you • Visible page — one keystroke • Full scrolling page — captured in segments and stitched, with sticky headers hidden so they appear once instead of repeating down the image Reach any of them from the keyboard, the toolbar icon, or a right-click. HOW THE NAME IS BUILT Cutline reads the page, not the pixels. A tool that inspects an image has to find a price somewhere in it and hope it is the product rather than an installment figure or a struck-through original. Cutline reads what the site itself declares. • Issue keys from Jira, Linear, GitHub, GitLab and Zendesk • Exception names from your selection or a stack trace • Product name and price from schema.org markup • HTTP error codes with their reason phrase • A cleaned page title, with unread counters, browser suffixes and marketing filler stripped out IT WOULD RATHER SAY NOTHING THAN GUESS A confidently wrong filename is worse than a timestamp, because it looks meaningful and is not. So an issue key found in body text must also appear in the page title. A denylist rejects lookalikes such as UTF-8 and RTX-4060. An exception name found in page text needs nearby stack-trace wording, so a documentation page about TypeError is not named after it — though a name you selected yourself is always trusted. A price is only used on a page that declares itself commerce. When nothing reliable is found you get a timestamp, which is at least honest. BEFORE IT SAVES A prompt shows the proposed name with a short countdown, then writes the file. Click it to edit, hover to pause, or discard. The image is already on your clipboard, so the wait costs you nothing. PRIVACY Cutline requests no host permissions. It cannot read pages in the background, cannot read other tabs, and never shows the "read your data on all websites" warning. It makes no network requests at all — no analytics, no accounts, no remote code, nothing to sign up for. Because filenames themselves can reveal things, banking sites, password managers, and any address containing /login, /checkout or /password are saved as a plain timestamp by default. Incognito windows always are. The list is yours to edit. WHERE FILES GO Downloads/Screenshots/2026-08/ by default. Date folders can be turned off, and you can add your own rules matching a hostname to a folder. Cutline will not guess a category for you — a misfiled screenshot is harder to find than a well-named one in a flat folder. Open source under the MIT License.