Skip to main content
A parser reads one kind of page on one site and returns the site’s own data as JSON. A Reddit search page comes back as posts, a YouTube video link as its transcript, a Google Maps search as places. There are 110 parsers.

Make a request

Send the page’s URL to scrape with a json format that names one parser. Get an API key in the Quickstart.

Read the response

The parser’s data is in data.json. The URL must be the kind of page the parser reads, or the call fails with invalidRequest. This real response is cut to the first items.

Options

Each parser has its own options. Send them in the options object of the format, like {"type": "json", "parser": "youtube-transcript", "options": {"language": "en"}}. A parser’s page lists the options it takes. Some parsers read more than one kind of page and list each as a view. Set options.view to pick one where the page says so. To read the next page of results, send the cursor from the response as the cursor option.

Choose a parser

Search and the web

Video

Social

Places and travel

Jobs

Shopping

Apps

Real estate

Ads

Companies and reviews

Price

Each parser has its own price per call, most of them $1 per 1,000 calls. 16 parsers work without an API key when you send no options, up to 10 calls a minute and 50 calls a day per IP address. See Rate limits. Paging and other options need an API key. See Pricing.

Read the list from code

GET /v1/parsers returns every parser. It is free and needs no API key.
Each entry has takes (the page it reads), an example URL, its options, the fields it returns, its price, keyless: true when it works without an API key, and a sample response. See List parsers.

Choose another format

  • JSON returns an object in the shape you describe, from any page.
  • Markdown returns the page text.
  • Scrape lists every format and option.