DFIRST AI provides built-in scraping tools that allow you to pull social media data—such as posts, comments, profiles, and engagement metrics—straight into your workflows for research and analysis.
Accessing the Scrapers
To begin extracting data, follow these steps:
Locate the central command dock at the bottom of your Whiteboard screen, and click on the Research icon, which is represented by a globe symbol.
Click the model selection dropdown to open the AI Agent menu, and choose the specific tool that matches your research needs.
Available Platform Scrapers
Choose the specific tool that matches your research needs. You can extract data from the following platforms (most extractions take approximately 10 seconds):
Facebook: Extracts posts, pages, and ads data from Facebook pages and profiles.
Instagram: Extracts posts, comments, stories, and profile details from Instagram profiles, hashtags, and locations.
LinkedIn: Scrapes LinkedIn profiles, company pages, posts, comments, and keyword-based post searches.
Reddit: Searches and extracts Reddit posts, communities, user profiles, and comments by keyword or URL.
TikTok: Scrapes TikTok videos, profiles, and hashtag feeds with sorting and pagination controls.
YouTube: Extracts video data, Shorts, and live streams from YouTube search results or direct URLs.
Once you select your scraper, enter your target URL or keyword in the prompt bar and press the arrow to run the extraction.
Evolving Your Strategy: Analyzing the Data
Scraping the data is only the first step. You can seamlessly route this raw data into a text model to generate actionable marketing insights.
Once your scraper node finishes generating the data on your whiteboard, select a Text model (such as GPT-5) from the dock.
Visually connect the output of your scraper node to your new text node.
Ask the AI to analyze the content. For example, you can prompt the AI to "turn these Instagram engagement insights into a sentiment analysis report" or "identify the top three best-performing posts and explain why they worked."
