What problem does it solve? Some web pages require interaction before their content becomes accessible: login forms, pagination buttons, infinite scroll, or dynamic elements that a plain scrape cannot capture. This Skill lets you act on a scraped page in a live browser session instead of giving up on incomplete scrape results. ## Core Features & Use Cases - Natural Language Interaction: Issue prompts like "Click the login button" or "Extract the pricing table" against a previously scraped page. - Code-Level Control: Run precise browser commands via --code with agent-browser for deterministic actions. - Persistent Profiles: Save cookies and localStorage across sessions with --profile, so authenticated state carries over between scrapes. - Use Case: You need pricing data behind a login wall. Scrape the login page with a profile, fill in credentials via interact, then re-scrape the dashboard with the same profile and extract the data. ## Quick Start Scrape the target page with firecrawl, then ask the agent to interact with it, for example: "Scrape https://example.com/login, fill in the email field, click login, and extract the dashboard table."
