textweb

# TextWeb Browser You have access to a text-based web browser via the `textweb_*` tools. Pages are rendered as structured character grids instead of screenshots. ## How It Works - `textweb_navigate(url)` — Opens a page and returns a text grid - `textweb_click(ref)` — Clicks element `[ref]` - `textweb_type(ref, text)` — Types into input `[ref]` - `textweb_select(ref, value)` — Selects dropdown option - `textweb_scroll(direction)` — Scrolls up/down/top - `textweb_snapshot()` — Re-renders current page - `textweb_press(key)` — Presses a key (Enter, Tab, etc.) - `textweb_upload(ref, path)` — Uploads a file to input ## Reading the Grid Interactive elements have reference numbers in brackets: | Element | Appears as | Action | |---------|-----------|--------| | Link | `[3]Click me` | `click(3)` | | Button | `[5 Submit]` | `click(5)` | | Text input | `[7:placeholder___]` | `type(7, "text")` | | Checkbox | `[9:X] Label` / `[9: ] Label` | `click(9)` to toggle | | Radio | `[11:●] Option` / `[11:○] Option` | `click(11)` | | Dropdown | `[13:▼ Selected]` | `select(13, "value")` | | File input | `[15:📎 Choose file]` | `upload(15, "/path/to/file")` | | Heading | `═══ TITLE ═══` | (not interactive) | ## Tips - The grid preserves spatial layout — elements near each other on screen are near each other in text - After clicking a link or submitting a form, you get the new page's grid automatically - Use `snapshot()` if you need to re-read the page after waiting for dynamic content - For multi-step forms, fill fields then click the Next/Submit button - Scroll down if you don't see what you're looking for — the initial view shows only the viewport