Bowmark

Bowmark gives agents a pre-computed cheatsheet for the task at hand. A cheatsheet is a compact, ready-to-run answer for one task on one site: a parameterized URL shortcut and/or a short ui_procedure of UI steps. Instead of burning tokens re-reading a site's DOM and guessing its way through the page, your agent calls ask({ site, task }) and gets the exact path back. It spends less, finishes sooner, and lands on the right action the first time.

Bowmark is free to use — no signup and no API key to manage. It's plain request/response REST — no websockets, no server-side browser, no async jobs. Your agent runs the cheatsheet in its own browser; Bowmark only supplies the path.

Methods. bowmark.ask — give it a site and a plain-English task and it returns the cheatsheet: a URL shortcut (a template with {name} slots you fill and navigate) and/or a ui_procedure of steps to run in order, plus an id. Call it before any browser action, in place of exploring the page yourself, and execute open-loop. bowmark.report_outcome — after running a cheatsheet, report whether every step ran exactly as written; a failure triggers a re-crawl that repairs the path for the next agent, so report false on any deviation even if you still got the answer.

Syntax & edge cases. site is a registrable domain, optionally with a product surface (google.com, docs.stripe.com, google.com/maps); task is intent, never a URL. Request the signed-in view with variants: { auth_state: "logged_in", role: "owner" } (also locale / region / currency). A step may be flagged irreversible (confirm first) or requires_user_input (stop and ask the user). Non-ok statuses: no_useful_data / site_not_supported → browse manually; ambiguous_scope → retry with scopeHint; rate_limited → back off until error.retry_after (only new cheatsheet synthesis is capped). Skip it for localhost, RFC1918 IPs, and open-ended search with no destination.