SaaS Template: YouTube Scraper

Generate structured data from YouTube videos using Vercel AI SDK generateObject function.

Only available in editions:

Enterprise

Warning

The YouTube scraping package (ytdl-core) works locally, but it will fail when deployed on services like Vercel or Fly.io due to YouTube blocking scrapers (Error Status: 410). To overcome this for TubeData, I had to request users' YouTube Cookies and X-YouTube-Identity-Token—an unconventional workaround that isn't included in the template. Additionally, TubeData automatically generates a smart schema from the scraped data, which can be updated in a CRUD-like interface. However, this feature is also not part of this template.

Features

YouTube Captions Scraper

Used the ytdl-core package to scrape YouTube captions.

Vercel AI SDK

Uses the 'generateObject' function to generate rows from the videos.

Feature Limits

Datasets are limited by the user's subscription plan.

References

YouTube scraping code

GitHub - Code

YouTube scraping demo

Vercel - Demo

SaaS MVP: Building a YouTube scraper from scratch

TubeData Launch: Extract data from YouTube videos

We respect your privacy. We respect your privacy.

TLDR: We use cookies for language selection, theme, and analytics. En savoir plus. TLDR: We use cookies for language selection, theme, and analytics. En savoir plus