Crawl4ai Interaction
AI-generatedSummary
Perform simulated web interactions of various types including element clicks, LLM-driven prompts, and authentication handling.
Inputs
- Interaction Type (required) — Choose the type of interaction to perform: 'element_click' to click on elements, 'form_fill' to fill and submit forms, 'llm_prompt' to automate interactions using a language model, or 'authentication' to handle login and session management.
- Element Configuration — When 'Element Click' is selected, specify the CSS selector of the element to click and optional waiting time after clicking (milliseconds).
- LLM Prompt Configuration — When 'LLM Prompt' is selected, provide the instructions text for the LLM, select the LLM provider (e.g., OpenAI GPT-4), and set the maximum tokens for the response.
- Authentication Configuration — When 'Authentication' is selected, provide the username, password, and optionally the login URL for performing authentication.
Output shape
a list of JSON objects
Each output item corresponds to an input item and contains the simulation result of the selected interaction type, including success status, details of the interaction (e.g., selector, prompt, username), a timestamp, and for errors an error message. The actual interactions are simulated; no real web interaction occurs in this implementation.
Examples
Example 1: Click a webpage element
Set Interaction Type to 'Element Click', specify a CSS selector and wait time after click. The node returns a simulated success response with interaction details.
Example 2: Use LLM to automate interaction
Set Interaction Type to 'LLM Prompt', provide prompt instructions, select LLM provider, and max tokens. The node simulates the prompt interaction and returns a placeholder LLM response.
Example 3: Authenticate to a web service
Set Interaction Type to 'Authentication', provide username, password, and optional login URL. The node simulates login and returns a simulated session ID in the response.