Firecrawl launches Directory to provide recruiting data for AI agents

Bringing professional data to AI

Hiring the right person often requires understanding specific past projects and experience rather than just looking at a job title. However, this professional information is frequently scattered across various company websites, databases, and platforms. Firecrawl is addressing this by introducing Directory, a new collection of data designed to help AI agents perform recruiting research.

Directory is part of Alexandria, which serves as a repository for Firecrawl's various data sources, including specialized indexes and live web content. The goal of Alexandria is to provide AI agents with the necessary information to perform useful tasks instead of relying only on their pre-existing training data. By adding professional profiles to this library, the platform aims to make recruiting a more streamlined research process.

The new collection includes more than 240 million professional profiles. These profiles contain employment histories and work contact information, which can be accessed through the Firecrawl plugin. This integration allows users to conduct deep research directly within popular AI interfaces like ChatGPT and Claude.

How the Directory search works

Directory pulls its data from several providers, including Apollo, FullEnrich, and Data Legion. These sources allow AI agents to search for people using a variety of filters such as job title, company, location, and specific keywords. This enables recruiters to move beyond simple keyword searches to find people based on the actual experience they possess.

When a user performs a search, the agent can retrieve information about a candidate's current role, previous employers, and professional background. For example, a search for senior backend engineers in New York can return hundreds of matches in a single free search. This provides a starting point for building a shortlist based on specific technical needs.

The system is designed to work within the flow of a conversation. Because it works inside ChatGPT and Claude, users can define a role, compare different backgrounds, and build a shortlist in the same chat window. This reduces the need to switch between multiple different tools to gather information on potential hires.

Moving from research to outreach

Finding a profile is only the first step in the recruiting process. Once a potential candidate is identified, the agent can use the web search functions of Firecrawl to find extra public details. This includes searching for technical writing, conference talks, open-source contributions, and previous interviews.

This extra layer of context helps recruiters understand a person's work beyond their listed employment history. It provides a better foundation for preparing specific questions to ask during an actual interview. By gathering this context, the AI helps turn a raw list of names into a more nuanced understanding of a candidate's capabilities.

Finally, the tool helps with the transition to outreach. Once a candidate is selected, the agent can look up available professional contact information. This allows recruiters to move from the identification phase to the outreach phase without manually searching for contact details across different platforms.

Access and pricing structure

Firecrawl has made Directory free to start for users who want to explore the data. Users can explore profiles and refine their searches without using any credits. This allows for a high level of exploration to see what kind of candidates are available in the database.

Credits are only used when a user chooses to retrieve detailed professional information, such as specific contact details. Every account comes with 1,000 free credits per month. This structure is intended to let users get started without needing an immediate additional subscription.

To use the tool, users can install the Firecrawl plugin in ChatGPT or Claude. For those who work with AI agents in a terminal, Firecrawl also offers skills that can be installed via a command line. These options ensure that different types of power users can access the professional data in the way that best fits their workflow.