Oya Browser

Read what the page logged: its own errors and warnings. Use this when a step failed, a page stalled, or the portal showed an error you cannot read on screen, it says what the page itself complained about.

List interactive elements on the page, with the same ids analyze_page gives. Lighter than analyze_page: no page text.

Read the requests the page made and how the server answered. Use failed_only when a submit or a handoff did not work: a 400 or 500 here names the request the server refused, which is the difference between bad data and a broken page. Judge only requests to the site you are working on, adverts, trackers and analytics fail on almost every commercial page and explain nothing about your task.

Keep a short note about how the site you are on works, for the next run that comes here: where a report lives, which menu hides a setting, a login or form quirk that cost you steps. Never task values, personal data or anything secret. Notes come back when a run reaches this site.

Ask for essential missing task constraints before dependent browser actions, or help with a login or other blocker. Group essential questions into one concise message and wait for the reply.

Call this right before you start the whole task again from the beginning, after an attempt went wrong. The playbook then keeps only the attempt that worked, so its replay does the task once. Do not call it to fix one field or step; fix that in place.

Finish the task by returning the data it asked for, in exactly this shape, with values quoted as the site writes them. Call it once, when you have everything; it ends the run. If the task cannot be done, reply FAILED: instead.

See the visible browser tab as an image. Use it when layout, images, icons or a canvas matter, or when analyze_page does not explain what is on screen. Element ids still come from analyze_page.

Scroll the page up or down.

Choose an option in a native dropdown (a select element from analyze_page) by the option text you see. Accepts {{placeholders}} and filters.

Start a new browser on this key and make it the one every other tool drives. Uses the key's configured provider unless one is given. It costs a session until stop_browser.

Stop a browser and release its session. Defaults to the browser the tools are driving.

Switch to one of your own tabs by ID: list_tabs marks them yours. Tabs the person or another agent is using are refused; use open_tab instead.

Type text into an input element by its ID (from analyze_page). Clears existing content first.

Write or update your plan: the steps this task needs, each marked done or not. Call it for any task of more than a few steps, always in the same turn as a page action, never alone: it does not touch the page.

Attach one of the task's files to an upload field. Only the files listed under FILES exist; you cannot upload anything else, and you never see or need the contents. The real <input type="file"> is normally hidden behind a styled "Choose file" / "Upload" / "Attach" button or a drop zone, so it often has no element id of its own: pass the id of the visible button, drop zone or field and the input behind it is found. Omit element_id only when the page has a single upload field.

Wait for an element matching a CSS selector to appear.

Wait until the page is ready: some text shows, the url contains something, and/or the network has gone quiet. With no text or url it waits for the network to settle. Use it after an action that loads results in the background, rather than analyzing again and again.