OpenErrand

Executes browser actions on behalf of the user, at their direction, in their own browser tab.

OpenErrand runs actions inside your own real browser, at your direction — log in somewhere, fill a form, upload a file, pull a report. Drive it yourself with a recorded recipe, or connect an app you trust and let it run the action for you, in your own logged-in session. Never a cloud browser. WHAT IT DOES • Opens a tab for an action you authorize and performs a fixed set of steps in it (navigate, click, fill, upload, wait, extract). • When an app is driving, it sends that app a minimized view of the page — interactive elements (labels and types, not values) — so it can decide the next step. Full screenshots and DOM are off by default and only sent if a signed recipe explicitly enables them. • Executes only a fixed, built-in set of native browser actions selected by incoming data. No remote code, no eval. YOU'RE IN CONTROL • Before an errand runs, you see exactly which sites and actions it's allowed — and you approve it. A recipe approved for one site can't touch another; the on-device enforcer checks every step against the recipe's signed list of allowed domains. • It only ever acts inside a tab it opened for an authorized errand — never your other tabs. • It connects only to apps you pair it with. The side panel lists every connection with a one-tap disconnect and a global kill switch. • Your logins are encrypted on your device (AES-GCM) and are never sent to OpenErrand or to the connected app — they go straight from your machine to the destination site. OPEN AND AUDITABLE The protocol and all security-critical code are open source (Apache-2.0). Anyone — including a security team — can audit the trust model and run a conformance suite against any server. Privacy policy: https://openerrand.app/docs/PRIVACY.html