Custom web datasets to build your competitive edge.

Turn any source of data on the web into a dataset delivered daily to your inbox.

Talk to us → Example datasets

Built by engineers from

Why use PointScrape

Consistently fresh data

PointScrape's intelligent prioritization ensures you're always getting the most up to date data.

Low cost & scalable

Intelligent prioritization also means lower compute costs and extreme scalability.

Use any data source

PointScrape's flexibility means any data type and source can be turned into a dataset, whether its HTML, JSON, or custom formats of product, news, or review data.

Our use cases

PointScrape data powers many ML and data capabilities:

LLM and NLP fine tuning

Adapting state of the art models requires significant in-domain datasets that PointScrape can help you build.

Owned-data enrichment

Enrich your owned data with context gathered from the web, like highlighting nearby highly rated restaurants or popular posts about your subject.

Competitive Intelligence

Gather intelligence on your market and competition, like reviews, prices, and availability for competitor products.

Some stats about us

121M

New rows added/month

New rows include news articles, reviews, products, product images, etc.

1.55 days

Median observed data age

PointScrape's intelligent prioritization ensures you are getting the most recent data, soon after it happens.

25,188

Dataset downloads on Kaggle

PointScrape datasets make up some of the most popular datasets on Kaggle.

The Process

Creating a dataset with PointScrape is a simple 3-step process:

Free Consultation

Provide your use case and examples of websites you’d like to scrape, and advise you on how the dataset can be created and a free quote for services.

Dataset Creation

Using your requirements, we set up PointScrape to build your dataset, with scheduled check-ins to make sure the produced dataset is just right.

Set & Forget

PointScrape continues gathering data after initial dataset creation, with a full suite of monitoring and alarms to keep it running.

Automated topic classification with Taggit

Taggit's text classification accurately identifies themes in scraped text. Integration is simple, and Taggit predictions help you quantify or identify topics critical to your business

Talk to us → Learn more

Schedule a Call

A short consultation is the first step to building your custom dataset.

Frequently Asked Questions

What is scraping?

Web scraping is the process of making requests to websites and saving the results. It is simple at a small scale, but quickly becomes more complicated as scale increases, as companies employ anti-scraping counter-measures, and as the data model grows complex. With PointScrape, you don't need to know about web scraping - just what data you want to turn into a dataset!

What if I'm not a dev?

That’s fine! You’ll work with an expert that can translate your needs into a working scraper, which will consistently deliver datasets to you with no code required!

Where can you get data from?

Anywhere! PointScrape is able to request and parse any kind of data that is publicly available. Even for complicated situations, PointScrape can handle it.

Can’t I just scrape the data myself?

For simple, small scale situations, you definitely should! But web scraping for even modest datasets get complicated quickly, and you should think about the following risks before deciding to build your own:

How do I avoid getting blocked by data sources when building and maintaining the dataset?
How do I handle data source changes to structure and layout in a backwards compatible way?
How do I figure out where new data is for larger scale sources?
How do I monitor and detect problems with the scraper before I have missed data that went uncollected?

It’s definitely possible, but not for the faint of heart! And even for capable engineers, PointScrape offers a cost effective solution that lets you move faster with lower total cost of ownership.

About Us

ThoughtVector has industry leading experience in NLP, machine learning engineering, and web-scraping to help you achieve all your data and machine learning needs.

"We've entered an unprecedented age of opportunity for data and machine learning. Don't let data moats hamstring your company's growth and market position."

Stuart Axelbrooke

Founder

Previously worked at