Runbook - Record it once, hand it over

Do the task once. Runbook writes the guide. Click through what you want to document and it produces numbered steps, a screenshot for each one with the control you clicked marked, and secrets already covered up — without a single file leaving your machine. WHY THIS ONE The tools that do this well upload every screenshot to somebody's servers. That is exactly why they are not allowed inside banks, hospitals, law firms, or any team with a compliance officer — and why the guides worth writing are the ones nobody can write. An admin panel, a payment console, a patient record system, an internal tool with real customer names on the screen. Runbook records in the browser, redacts in the browser, and exports to a file. There is no account, no sync, no link, and no server to be breached, because there is nowhere for anything to go. IT HIDES SECRETS BEFORE THEY ARE STORED An API key, a card number, an IBAN, an email address, a bearer token, a private key, a one-time code. Seventeen kinds of secret are recognised while you record, and the screenshot is masked before it is written to disk — not when you export it. A key sitting in a database from Tuesday to Friday is a key that leaked on Wednesday. Fields the page itself marks as secret are never read at all. The guide says "Type your password"; it does not say what you typed. IT WRITES A DOCUMENT, NOT A TRANSCRIPT Thirty keystrokes are one step. A double click is one step. The click that only put the cursor in a box is not a step at all — "click the email field, then type your email" is how a machine describes it, and nobody finishes reading that. Every step is a sentence you can rewrite, and your words are never regenerated over. IT IS HONEST ABOUT ITS OWN PICTURES Chrome limits how often an extension can take a screenshot. When two clicks land inside that limit and the screen has not moved, the step reuses the previous frame — and says so, in the caption. It never reuses a frame after the page scrolled or navigated, because a marker on the wrong screen is worse than no picture. EXPORT - Markdown, to paste into a ticket, a wiki or a pull request - HTML, one self-contained file that opens anywhere, forever - PDF, printed by Chrome itself, with selectable text and working links PERMISSIONS One site at a time, asked for when you start recording there. Installing shows no warning about reading your data on all websites, because it does not ask for that. FREE - 3 guides - 15 steps per guide - Markdown and HTML export, with no watermark PRO - USD 29 once, no subscription - 500 guides - 300 steps per guide - PDF export with a cover page and your logo - reusable redaction profiles WHAT IT WILL NOT DO It does not record video, because a video cannot be searched, translated, diffed or pasted into a ticket. It does not offer a shared team library, because that needs a server and the server is the thing being avoided. It does not write your step descriptions with a language model, because that means either sending your screen somewhere or being confidently wrong about buttons it cannot read.