Skip to content
Datasets · early access

Web scraping

Collect rows from one or many pages in a real browser and keep them up to date. Monity tells you which rows are new, changed or gone, and hands them to your spreadsheet, your code or your AI assistant.

Ask for early access Read the docs
  • Pick one item, get the whole list
  • Alerts on new, changed and removed rows
  • CSV, JSON, Markdown and Excel
Making a dataset in Monity: one click on an article of the Monity blog outlines the articles like it, with Collect these

Rows from many pages

Click one item. Get rows from every page.

Open the page in the steps editor and click one item of the list: Monity finds the others like it and proposes the fields. Add Next page and For each item, check the rows with Run to here, and save. Every run replays exactly those steps.

  • One address or up to 1,000, a CSV, or pages found from a sitemap
  • Searches, filters, pages of results and each item’s own page
  • Fields cleaned on every run: dates, numbers, prices, addresses
  • Monity AI can draft the steps from one sentence, for you to check
Make a dataset
The Rows tab of a dataset of the Monity blog: a search box, the All rows, New and Changed filters, and the articles with their title, date and link

What changed

Every run says which rows are new, changed or gone.

Monity compares each run with the one before, row by row. The Changes tab sums up what moved in a sentence and lists the rows, with the old and the new value of every field that changed.

  • New, Changed, Gone and Back
  • Search the rows, or show only the new and changed ones
  • Every run kept, with the pages it read and the credits it used
Rows, changes, pages and runs
A dataset run’s record: 29 rows from 4 list pages for 4 checks, read in full, with what each step did and a screenshot after it

Alerts

Hear about new rows where your team already works.

A dataset collects again on its own schedule, from every hour to once a month, and alerts on the rows that moved. Choose which count: new rows, gone rows, or changes in only the fields you pick, such as the price.

  • Email, Slack, Microsoft Teams, Discord, Telegram, Google Sheets, webhooks and push in the mobile app
  • Google Sheets: a sheet row for each new, changed or gone row
  • Webhooks: the rows and their changes as JSON
  • Send test shows a sample alert before the first real one
Dataset schedule and alerts
A dataset’s alerts: tell me when rows are New, Changed or Gone, and the channels to send them to

Out of Monity

Take the rows to a spreadsheet, your code or your AI assistant.

Export every row as CSV, JSON, JSON Lines, Markdown or Excel. Or connect Claude Code, Codex or another MCP client with an API key, and ask about your datasets in the conversation.

  • Values that would run as a spreadsheet formula are written as text
  • Rows as data or as a ready Markdown table, searchable and paged
  • What a run found new, changed or gone, for the assistant to explain
Datasets in Claude and Codex

Priced in credits

The same credits as your monitors. Setting up is free.

A run spends the workspace’s credits: 1 for each page it opens. Run code is free, and an AI read step costs 5 more for each page it reads. A list of 4 pages read every day is about 120 credits a month.

  • The cost of a run and of a month, on Settings
  • Each run’s credits on the Runs tab
  • Paid from the plan’s monthly credits, the same ones your tasks use
Credits for datasets
A dataset’s Cost card: the credits for its list pages and AI read, what a run costs and what that comes to in a month

What the alerts say

Set the rows you need once. Get alerts about the ones that moved.

You set the steps once. Each alert names the dataset, counts what moved and lists the rows.

  • Every open role on the careers pages of Relaylane, Coastal CRM and Fieldnote: title, team, location and link
    Slack · #hiring-watch
    Competitor hiring: 4 new, 1 gone Enterprise Account Executive, London · Relaylane careers, and 3 more.
  • The price and stock of every robot vacuum on these three shops, every 6 hours. Alert only when a price changes
    Email · pricing@company.com
    Robot vacuum prices: 2 changed Northwind S8: price $499 → $449 · northwind-shop.com
  • Each new product Fieldnote and Coastal CRM add to their integrations pages
    Google Sheets · Competitor integrations
    Competitor integrations: 3 new Three sheet rows added, each with the time, New, the page and its fields.
  • Every listing under $500,000 on these two property sites, every hour
    Telegram · private chat
    Listings under $500k: 2 new, 1 changed 2 bedrooms, Harbour Street: price $489,000 → $469,000

Relaylane, Coastal CRM, Fieldnote and Northwind are example companies.

FAQ

Questions about web scraping with Datasets

What is a dataset in Monity?
Rows Monity collects from one or many pages in a real browser, such as the listings on several sites or the products of a set of competitors. Monity collects them again on a schedule, compares each run with the one before and tells you which rows are new, changed or gone.
How is it different from a scraping API?
There is nothing to code. You click one item on the page, Monity finds the rest, and you see the rows with Run to here before anything is scheduled. Then Monity replays the same steps on a schedule and alerts on what changed. The rows come out as files, into Google Sheets or a webhook on every run, or to Claude and Codex over MCP. Monity has no scraping REST API.
How many pages can a dataset read?
A dataset starts from up to 1,000 addresses, or from the pages Monity finds from a site’s sitemap and links (up to 1,000 an address). On each, the steps can search, filter, read up to 100 pages of results and open each item’s own page. A run has limits you set: up to 100 list pages an input, 10,000 rows and 60 minutes.
How often does it collect?
On the schedule you choose: every hour, every 3, 6 or 12 hours, every day, every week or every month. Run now collects at once.
What does it cost?
Datasets use your plan’s credits, 1 credit = 1 check: 1 for each page a run opens in a browser, 5 more for each page an AI read step reads, and 2 a minute while a run uses a saved login. Run code is free, and so is trying the steps with Run to here. See credits for datasets.
Can it read pages behind a login?
Yes. On the dataset’s Settings, choose a saved login, or sign in once in the Live Browser, and every run signs in with it before its first step. Passwords are never kept in the steps.
What about sites that block bots?
Monity reads every page directly, in a real browser. When a site shows a security check instead of its content, the page is marked Blocked and you can read it through your own proxy, added once in Integrations. Datasets never use Monity’s pooled proxies, and Monity does not promise to get past every site’s bot protection.
Can Claude or Codex use my datasets?
Yes. Connect Claude Code, Codex or another MCP client to Monity with an API key; it can list your datasets, read their rows as data or as a Markdown table, see what changed and start a run. See Datasets in Claude and Codex.
How do I get Datasets?
Datasets are in early access. In the app, open Datasets and click Ask for early access, or write to us with the pages you want covered, and we switch them on for your account.

Related

Tell us what you want collected.

Datasets are in early access. Write to us with the pages or the market you want covered, and we switch them on for your account.

Ask for early access Start free
  • The same steps on every run
  • Paid from your plan’s credits
  • Every alert channel
A Collect list step in the steps editor: the fields of each article with their transforms, samples and the row key
Kamil, Alex and Martyna Median reply under 6 hours

Have a question?

Pricing, security, or scoping - pick a route and you will land in a real person's inbox.

S Talk to sales Pricing and scoping H Get support Help with a monitor D Book a demo 30-minute walkthrough