ADKTabInfo represents information about a browser tab.

Agent is the main interface for browser automation with LLM.

AgentConfig configures the browser agent.

AgentHistory manages the execution history of an agent.

AnnotatedOptions extends Options with annotation settings.

AnnotationConfig configures how annotations are drawn.

BoundingBox represents an element's position and size on the page.

BoundingBoxInfo provides bounding box data for annotation.

BoundingBoxInfoInterface defines the interface for bounding box info. This matches the screenshot.BoundingBoxInfo interface.

Browser wraps rod.Browser with enhanced functionality.

BrowserAgent is the main agent that controls browser automation via LLM using ADK.

BrowserToolkit holds browser context for tool execution.

ClearAndTypeArgs is the input for the clear_and_type tool.

ClearAndTypeResult is the output for the clear_and_type tool.

ClickArgs is the input for the click tool.

ClickResult is the output for the click tool.

CloseTabArgs is the input for the close_tab tool.

CloseTabResult is the output for the close_tab tool.

Config holds agent configuration.

Config holds browser configuration.

Config holds test runner configuration.

ContextualAgent wraps Agent with a context for convenience methods.

DoneArgs is the input for the done tool.

DoneResult is the output for the done tool.

DoubleClickArgs is the input for the double_click tool.

DoubleClickResult is the output for the double_click tool.

Element represents an interactive page element.

ElementInfo provides element data needed for annotation.

ElementInfoInterface defines the interface for element info used in screenshot annotations. This matches the screenshot.ElementInfo interface.

ElementMap holds all interactive elements on a page.

ElementMapAdapter adapts dom.ElementMap to screenshot.ElementMapInterface.

ElementMapInterface defines the interface for element maps used in annotations. This allows the screenshot package to work with element maps without c

EvaluateJSArgs is the input for the evaluate_js tool.

EvaluateJSResult is the output for the evaluate_js tool.

Expectations defines what to validate after task completion.

ExtractContentArgs is the input for the extract_content tool.

ExtractContentResult is the output for the extract_content tool.

Extractor handles DOM element extraction from a page.

FocusArgs is the input for the focus tool.

FocusResult is the output for the focus tool.

GetPageStateArgs is the input for the get_page_state tool (no args needed).

GetPageStateResult is the output for the get_page_state tool.

GoBackArgs is the input for the go_back tool.

GoBackResult is the output for the go_back tool.

GoForwardArgs is the input for the go_forward tool.

GoForwardResult is the output for the go_forward tool.

HistoryItem represents a single step in the agent's execution history.

HoverArgs is the input for the hover tool.

HoverResult is the output for the hover tool.

ListTabsArgs is the input for the list_tabs tool (no args needed).

ListTabsResult is the output for the list_tabs tool.

MessageManager handles conversation state and message construction for the LLM.

MessageManagerConfig configures the message manager.

---- Tool Argument Structs (ADK format with json + jsonschema tags) ---- NavigateArgs is the input for the navigate tool.

NavigateResult is the output for the navigate tool.

NewTabArgs is the input for the new_tab tool.

NewTabResult is the output for the new_tab tool.

Options configures screenshot capture.

GetPageInfo returns information about the current page state.

Preset defines token/quality tradeoffs for different use cases.

ReloadArgs is the input for the reload tool.

ReloadResult is the output for the reload tool.

Result represents the outcome of a task execution.

Result represents the outcome of an agent run.

ScreenshotArgs is the input for the screenshot tool.

ScreenshotResult is the output for the screenshot tool.

ScrollArgs is the input for the scroll tool.

ScrollResult is the output for the scroll tool.

ScrollToElementArgs is the input for the scroll_to_element tool.

ScrollToElementResult is the output for the scroll_to_element tool.

SendKeysArgs is the input for the send_keys tool.

SendKeysResult is the output for the send_keys tool.

SensitiveDataFilter filters sensitive data from messages.

SerializeOptions configures how elements are serialized.

StealthConfig configures anti-detection measures.

Step represents a single action in the execution sequence.

Step represents a single step in the agent's execution.

SwitchTabArgs is the input for the switch_tab tool.

SwitchTabResult is the output for the switch_tab tool.

TabInfo contains information about a browser tab.

TabInfo contains information about an open tab.

TestCase represents a single test definition.

TestResult holds the outcome of a single test.

TestSuite represents a collection of tests from a YAML file.

TokenBudget helps manage token allocation across message components.

TokenCounter provides token estimation for text content. This is a simple approximation - actual token counts depend on the specific tokenizer.

TypeTextArgs is the input for the type_text tool.

TypeTextResult is the output for the type_text tool.

Viewport defines browser viewport dimensions.

WaitArgs is the input for the wait tool.

WaitResult is the output for the wait tool.

elementInfoAdapter adapts dom.Element to screenshot.ElementInfo.

extractionResult is the structure returned by the extraction JavaScript.

presetConfig defines the configuration for each preset.