How to Scrape 10,000 Leads with Claude Code

TL;DR

Install the ScrapeGraphAI GitHub repository through Claude Code, add an OpenAI API key with payment enabled, and describe the businesses you want to collect. The demonstrated workflow produced more than 9,000 accounting-related leads in a CSV after running for under an hour, while the presenter reported spending about $2 instead of paying for a monthly lead database.

Transcript

In this video, I'm going to show you how you can get up to 10,000 leads in any niche or industry that you want with just one prompt. Okay, so watch this, guys. So, I can just type in Claude here. Find me 10,000 accounting leads in the USA and let that run. We're going to wait a couple minutes. And here is the output. So, you can see here a full lis... Read More

Key Insights

• Claude Code can install ScrapeGraphAI from its GitHub repository after receiving the repository URL and a plain-English installation request. The repository already contains the scraping instructions, so the demonstrated workflow relies on Claude Code to interpret those instructions and guide the local setup.
• An OpenAI API key is used as the language-model credential for the scraping workflow. The presenter creates it through platform.openai.com, gives it a name, copies it for later use, and enters it when the ScrapeGraphAI setup opens an API-key prompt.
• A lower-tier OpenAI model is selected because lead scraping is described as a relatively simple task. The presenter says OpenAI's lower-cost option is cheaper than Claude's available alternatives for this purpose, which is why Claude Code is given an OpenAI key.
• API billing must be enabled before the scraper can complete its work. The first attempt reports that the OpenAI account has no payment attached, so the presenter returns to the account, adds $5, and then continues the configured scraping process.
• ScrapeGraphAI can collect up to 10,000 leads in one run, according to the presenter. The example request targets accountants, bookkeepers, and tax preparation services throughout the United States, while asking the system to find as many relevant businesses as possible.
• The scraping request can be written in everyday language rather than code. After confirming that setup is complete, the presenter tells Claude Code which business categories and country to target, then states that 10,000 results would be desirable.
• The completed job returns its lead records as a CSV spreadsheet. In the demonstration, the output contains more than 9,000 different leads and is described as finishing in under an hour, although the suggested general waiting range is approximately 30–90 minutes.
• The demonstrated cost is about $2 for roughly 10,000 leads, according to the presenter. She contrasts that amount with lead-finder platforms that she says typically cost about $99 per month and provide 10,000 leads, making the self-run workflow cheaper in her example.
• More videos from this channel:
  • How to Build Viral AI Videos with Claude
  • How to turn AI automations into a scalable micro SAS?
  • How to use AI for personal tutoring and reports
• More from this channel: Helena Liu

Questions & Answers

Q: How do you scrape leads with Claude Code and ScrapeGraphAI?

Give Claude Code the ScrapeGraphAI GitHub repository URL and ask it in plain English to help install the repository for lead scraping. Follow the additional setup instructions it returns, enter an OpenAI API key when prompted, and confirm that payment is enabled on the API account. Then open a new window and describe the industries, location, and approximate number of leads you want.

Q: How do you install ScrapeGraphAI using Claude Code?

Open the desktop version of Claude Code in a new window, provide the ScrapeGraphAI GitHub URL, and ask Claude Code to install the repository so you can scrape leads. Claude Code reads the repository instructions, performs the local installation, and supplies any additional commands that must be followed. During the demonstrated setup, it also opens a prompt where the API key can be entered.

Q: Why does the workflow use an OpenAI API key inside Claude Code?

The presenter describes lead scraping as a relatively simple task that does not require an expensive language model. Based on her comparison of the available lower-tier models, she selects an OpenAI option because it has a lower token cost than Claude's offerings for this use case. Claude Code therefore manages the workflow while the OpenAI key supplies the model access used by ScrapeGraphAI.

Q: How do you create the OpenAI API key needed for lead scraping?

Go to platform.openai.com, log in, open the API keys section, and create a new key. Give the key a name, copy it, and store it somewhere so it is available during the ScrapeGraphAI installation. The demonstrated account initially lacked payment information, so the presenter added $5 before the scraping workflow could continue successfully.

Q: What prompt can you use to find accounting leads in the United States?

After completing the ScrapeGraphAI setup, tell Claude Code that the setup is finished and request as many accountants, bookkeepers, and tax preparation services in the United States as possible. The presenter includes a target of 10,000 leads in the request but also tells the system to find as many as it can. The instructions are written in ordinary conversational English.

Q: How many leads can ScrapeGraphAI collect in one run?

ScrapeGraphAI can process up to 10,000 leads at once, according to the presenter. The accounting example did not return the full maximum, but it produced more than 9,000 different leads. The requested categories included accountants, bookkeepers, and tax preparation services from across the United States, and the resulting records were delivered in a CSV spreadsheet.

Q: How long does the Claude Code lead-scraping process take?

The presenter recommends allowing approximately 30–90 minutes for a large scraping request to finish. She suggests starting the process before bedtime and returning later to a completed spreadsheet. In the accounting example, more than 9,000 leads were collected in under an hour, although the exact duration can only be stated for the demonstrated run.

Q: How much does it cost to scrape 10,000 leads with this workflow?

The presenter reports spending about $2 to scrape roughly 10,000 leads and says the expected cost is around $1 or $2. She initially added $5 to the OpenAI account after the setup reported that no payment was attached. She compares this with lead-finder platforms that she says commonly charge about $99 per month for 10,000 leads.

Summary & Key Takeaways

• The workflow begins by creating an API key at platform.openai.com, copying it, and storing it securely for setup. The presenter chooses a lower-tier OpenAI model because the scraping task is relatively simple and, according to her comparison, its token cost is lower than the available Claude options for this particular use case.

• Claude Code handles installation through a plain-English request containing the ScrapeGraphAI GitHub URL. It reads the repository instructions, installs the necessary components on the local machine, provides additional commands to follow, and opens a place to enter the API key. The presenter enters an OpenAI key when another provider's key is requested.

• After setup, the user opens a new Claude Code window and requests accountants, bookkeepers, and tax preparation services across the United States. ScrapeGraphAI can process up to 10,000 leads at once, according to the presenter. The job runs for roughly 30–90 minutes and returns the collected records in a CSV spreadsheet.