Turn Email Lists into Rich Company Datasets

Grok 4 Fire Enrich is an open-source data enrichment tool that transforms email lists into detailed company datasets with profiles, funding information, tech stacks, and more.

This multi-agent tool uses Firecrawl for web scraping, Grok 4 for agent orchestration, and GPT-4o for intelligent data extraction.

GitHub Repo

Deploy

Features

Multi-Agent Architecture: Five specialized agents work sequentially: Discovery, Company Profile, Financial Intel, Tech Stack, and General Purpose. Each building on previous findings for maximum accuracy.
Real-time Processing: Watch your data get enriched row-by-row through Server-Sent Events with live updates.
Drag & Drop CSV Upload: Simple interface accepts standard CSV files with email addresses.
Customizable Data Fields: Choose from pre-defined options (industry, funding stage, tech stack, CEO) or create custom fields using natural language.
Source Citations: Every piece of extracted data includes the URL where it was found, ensuring complete transparency.
Smart Email Filtering: Automatically skips personal email providers (Gmail, Yahoo, etc.) to focus on corporate data.
Unlimited Mode: When run locally, removes demo restrictions for unlimited rows and columns.
Type-Safe Architecture: Built with Zod schemas for reliable data validation and easy extensibility.

Use Cases

Sales Teams: Enrich prospect lists with company size, industry, funding stage, and key personnel to improve targeting and personalization.
Marketing Research: Build competitor analysis datasets with tech stacks, funding information, and market positioning.
Investor Due Diligence: Gather startup information, including funding rounds, employee counts, and technology infrastructure for investment research.
Business Development: Identify potential partners by analyzing company profiles, headquarters locations, and business models.
Lead Qualification: Automatically score and categorize leads based on company size, industry, and growth stage indicators.

Case Studies

Example: Enriching [email protected]

Input Data:

{
  "email": "[email protected]"
}

Process Flow:

Discovery Agent extracts domain (firecrawl.dev) and identifies it as a B2B SaaS company
Company Profile Agent classifies industry as “Developer Tools / Web Scraping APIs”
Financial Intel Agent finds funding stage as “Seed” and connection to Mendable AI
Tech Stack Agent analyzes GitHub repos to identify Node.js, Python, Redis, and Kubernetes
General Purpose Agent locates CEO information: Eric Ciarla, Co-founder

Final Output:

{
  "email": "[email protected]",
  "industry": "Developer Tools / Web Scraping",
  "ceo": "Eric Ciarla, Co-founder and CEO",
  "fundingStage": "Seed (Part of Mendable AI)",
  "techStack": "Node.js, Python, Redis, Kubernetes",
  "sources": [
    "firecrawl.dev/about",
    "linkedin.com/company/firecrawl",
    "github.com/mendableai/firecrawl"
  ]
}

Installation

1. Get API Keys:

Firecrawl API key from firecrawl.dev/app/api-keys
Grok 4 API key from x.ai/api
OpenAI API key from platform.openai.com/api-keys

2. Installation:

   git clone https://github.com/mendableai/grok-4-fire-enrich
   cd grok-4-fire-enrich
   npm install

3. Create .env.local file:

   FIRECRAWL_API_KEY=your_firecrawl_key
   GROK_API_KEY=your_grok_4_key
   OPENAI_API_KEY=your_openai_key

4. Run the Application:

   npm run dev

5. Open http://localhost:3000 in your browser to start using the tool.

Usage

Upload CSV: Drag and drop a CSV file containing email addresses into the interface
Select Fields: Choose data points to extract (industry, funding stage, tech stack, CEO, etc.) or add custom fields
Start Processing: Click “Enrich Data” to begin the multi-agent extraction process
Monitor Progress: Watch real-time updates as each row gets processed through the five-phase system
Export Results: Download the enriched dataset with source citations

Pro Tip: For best results, use corporate email addresses rather than personal ones. The system automatically detects and skips Gmail, Yahoo, and other consumer email providers.

Pros

Cost-Effective: Free alternative to expensive commercial data enrichment services that can cost hundreds per month.
Complete Transparency: Every data point includes source URLs.
Highly Customizable: Add custom fields using natural language descriptions beyond the standard options.
Open Source: Full access to code for modifications, improvements, and custom integrations.
Real-time Processing: Live updates show progress without waiting for batch completion.
Smart Context Building: Sequential agent execution means later phases benefit from earlier discoveries.
Local Control: Run entirely on your infrastructure without sending data to third-party servers.

Cons

API Dependency: Requires three separate API keys (Firecrawl, Grok 4, OpenAI).
Processing Speed: Sequential agent execution, while more accurate, takes longer than simple parallel processing.
Accuracy Variations: Results depend on publicly available information and web scraping success rates.
Setup Complexity: Requires technical setup, including API configuration and local development environment.

Related Resources

Firecrawl Documentation: firecrawl.dev/docs – Learn about the web scraping API that powers data collection
Grok 4 API Guide: x.ai/api – Official documentation for the AI model handling agent orchestration
OpenAI API Reference: platform.openai.com/docs – GPT-4o integration details for data extraction
Zod Schema Validation: zod.dev – Type-safe schema system used throughout the application
Next.js 15 Documentation: nextjs.org/docs – Framework powering the user interface
CSV Data Processing Best Practices: papaparse.com – Library used for robust CSV parsing

FAQs

Q: Is Grok 4 Fire Enrich completely free?

A: The tool itself is open-source and free to use. However, you will need to pay for the API usage of Firecrawl, Grok 4, and OpenAI, depending on your usage.

Q: What types of data can Grok 4 Fire Enrich extract?

A: The tool can extract company names, industries, headquarters locations, founding years, employee counts, funding stages, total funding raised, tech stacks, CEO information, and custom fields you define. Each data point comes with source citations showing where the information was found.

Q: How accurate is the data extraction?

A: Accuracy depends on publicly available information and the multi-agent architecture. The sequential approach where each agent builds on previous discoveries typically produces more accurate results than parallel processing. However, some companies may have limited public information or outdated web presence, which can affect completeness.

Q: Can I process large datasets?

A: The public demo limits processing to 15 rows and 5 columns, but running locally enables unlimited mode. Processing speed depends on your API rate limits and the complexity of data being extracted. Most users can process 50-100 emails per hour.

Q: Can I extend the tool with custom agents?

A: Yes, the architecture supports custom agents. You can modify existing Zod schemas in the /lib/agent-architecture/agents/ directory or create new agents by implementing the AgentBase interface. The field routing system automatically categorizes requests to appropriate agents.

Q: Is my data secure?

A: When run locally, all processing happens on your infrastructure. The tool only sends web scraping requests to Firecrawl and AI processing requests to OpenAI/Grok 4. Your original CSV data never leaves your system unless you choose to export results.