I offer cold email personalization services and need to scrape leads from YouTube channels (email + channel link). Usually I scrape via website, but for this client I must use YouTube. How can I scrape YouTube pages and then use ChatGPT for AI cold email personalization?
He suggests several simple methods: use a marketplace like Appify to find a YouTube channel scraper, or scrape YouTube videos in a niche, extract channel info from those videos, deduplicate, then scrape each channel page using a YouTube channel scraper tool (e.g., by providing a channel ID). Once you have channel data, you can parse for emails; if missing, use a second route: look for the channel’s website and run it through an email finder/verification service, which will find an email 40‑50% of the time. Then you can send personalized messages via ChatGPT.
To find thousands of YouTube creators with emails, start by searching YouTube for a relevant keyword (e.g., 'automation'). Filter the results by view count—say, only channels with 100K+ views. Grab the channels from those videos, then feed those channel lists into an email scraper (such as Apify/Ampify) to pull out the associated email addresses. Those emails can then be loaded into your outreach platform (like Instantly) for your campaign selling content ideas or training video systems.
Congratulations on landing $2,500—that's more than 90% of the world makes outside traditional employment. You should be stoked. To build the search intent video system, start by scraping the simplest source: LinkedIn jobs. Use an Appify actor (like a LinkedIn job scraper) to pull job postings. You don’t need company details at first; just scrape the jobs. Then, to get emails, scrape or enrich using the company name to find the website, and use a mailfinder tool (any mailfinder) to get emails from those domains. If the mailfinder doesn’t give you the URL, you’ll need to scrape company details to get the website. Once you have the websites, enrich them to find people at those companies, identify decision makers, and then feed the email and domain into an AI tool that searches the domain and combines results with the email to pull out the person's name (e.g., turning 'nick@leftclick.ai' into 'Nick Sarif'). This makes your outreach far more effective. After gathering the emails, customize your outreach using the scraped information, and you can also scrape job details if needed. This whole process lets you turn raw data into personalized, high‑conversion cold emails.
Ideal customers are creators who are just starting to standardize their content, have around 10K–100K followers, are at the early stage of exponential growth, and are looking to outsource or work with a team to reduce costs. You can serve them by using a transcript to automatically generate posts for ten social platforms—either fully automated or with a person handling it. To find these leads, scrape Instagram, Twitter, and YouTube for channels in that follower range. For example, use a YouTube scraper with about fifty search terms to get roughly a thousand videos for $5, which corresponds to about six hundred creators. Take the top ten percent (about fifty to a hundred channels), sort them by subscriber count, and then use a YouTube channel scraper to pull contact details from video descriptions—Instagram, website, etc.—giving you Instagram, Facebook, X/Twitter, and LinkedIn profiles. With those contacts, take one of their long‑form YouTube videos and use your system to create customized content for each platform at a marginal cost of just a few cents. Finally, run personalized outreach based on the video or another asset, ensuring each message includes some level of customization.
Don't personalize the whole email—personalize only small snippets using enhanced variables. For example, use a template like 'Hey {first name}, love {thing we have in common}, fun to see you mention that on {platform} where I got it.' Then continue that vein to your offer. If AI writes the whole email, it will likely produce something stupid and obvious as AI‑generated. Keep the AI constrained everywhere and feed it only small variables (enhanced variables) with examples, letting it spin the initial info into a tightly woven, human‑sounding message. For research, use Claude to look up the person, compile company info, and grab all social media profiles; if you can't, use scrapers or similar tools.