Scrape any company website. One call.
Your agent sends a domain. MoltSets fetches the homepage and returns structured JSON: headings, key pages, emails, phone numbers, social profiles and the technologies it runs on.
POST /tools/scrape_website { "domain": "acme.com" }
200 OK · status: ok · 1 scrape token
---
title Acme · Billing software for SaaS teams
key pages pricing /pricing · careers /careers · demo /demo · docs null
emails sales@acme.com
phones +1 415 555 0100
socials linkedin · x · github
tech hubspot · stripe · intercomOne domain in, the homepage as data out
page content
What the company says about itself, in its own words.
- Title, description and Open Graph text
- H1–H3 headings
- A text excerpt of the visible copy, and counts
key pages
The pages you’d otherwise go hunting for.
- Contact, about, pricing, careers and demo
- Blog, docs, privacy, terms, login and signup
- Navigation and footer links
contact points
Ways to reach the company, as published on its homepage.
- Emails found on the page
- Phone numbers found on the page
- Social profiles
technologies
The tools a company runs its website on.
- Detected technologies
- Read from the homepage itself
- Returned with everything else, in the same call
Skip the HTML. Get the fields.
scrape_website never returns raw HTML or images. Your agent gets structured JSON it can use straight away, not thousands of lines of markup. Read status, not the HTTP code, to tell a hit from a miss.
- 01
domain: "acme.com"Pass a domain. A full URL is reduced to its host. - 02
homepageThe homepage is always the page fetched. - 03
status: "ok"Structured JSON back, for 1 scrape token. - 04
status: "pending"The site blocked the direct fetch. Ask again afterretry_after_seconds. Free. - 05
status: "not_found"No data for that domain. Free.
Blocked or empty? Not charged.
You spend one scrape token per domain served, whether it’s fetched now or served from the 30-day cache. Anything that doesn’t come back costs nothing.
- A site that blocks the direct fetch returns pending with a retry delay. Ask again after it.
- A domain with no data returns not found.
- Pending and not-found results cost nothing.
- Scrape tokens are a separate balance from your regular tokens.
What agents scrape for GTM
01 / Account qualification
Qualify accounts from what their website says: the headline, the headings and the copy on the homepage.
02 / Page discovery
Find the contact, pricing, careers and demo pages without crawling. A page that isn’t there comes back null, never guessed.
03 / Tech detection
Detect the technologies a company uses, then route or score accounts by stack.
04 / Personalisation
Hand homepage copy to your AI agent and write one-to-one openers at scale.
An add-on to a paid Standard Subscription
Scrape tokens sit on top of a paid Standard Subscription. They aren’t available on the Free plan. Pick your plan first, then add the scrape token subscription that fits your volume.
- Free: scrape tokens not available
- Solo, $27/month: add scrape tokens
- Pro, $97/month: add scrape tokens
- Team, $397/month: add scrape tokens
- Max, $997/month: add scrape tokens
Scrape tokens from $0.142 per 1k
One token per domain, including results served from the 30-day cache. The more you subscribe to, the less each 1k costs.
- Charged per domain served
- Pending and not-found results are free
- Requires a paid Standard Subscription
| tokens / month | volume | price |
|---|---|---|
| 37,500~1,250 / day | $19/moHobby · $0.507 / 1k | |
| 100,000~3,333 / day | $29/moStarter · $0.290 / 1k | |
| 200,000~6,667 / day | $49/moDeveloper · $0.245 / 1k | |
| 500,000~16,667 / day | $99/moGrowth · $0.198 / 1k | |
| 875,000~29,167 / day | $149/moBusiness · $0.170 / 1k | |
| 1,750,000~58,333 / day | $249/moScale · $0.142 / 1k |
Costs and outcomes are in the Scrape Website docs.
Website scraping questions
Every parameter, status and error code is in the Scrape Website docs.
Each domain served costs 1 scrape token, whether it’s fetched now or served from the 30-day cache. pending and not_found results cost nothing. Scrape tokens are a separate balance from your regular tokens; get_billing shows both.
No. Scrape tokens are an add-on to a paid Standard Subscription, from Solo upwards. Upgrade from Free first, then add the scrape token subscription that fits your volume. See pricing.
No. The homepage is always the page fetched. Pass a domain such as acme.com; a full URL like https://www.acme.com/about is reduced to acme.com. To reach deeper pages, use the key page links that come back.
Structured JSON: title, description and Open Graph text, H1–H3 headings, navigation and footer links, key pages, emails, phone numbers, social profiles, detected technologies and a text excerpt of the visible copy. It never returns raw HTML or images.
A site that blocks the direct fetch may first return status: "pending" with retry_after_seconds. Ask again after that delay. Pending results cost nothing.
Read status, not the HTTP code. A domain with no data returns HTTP 200 with status: "not_found", and costs nothing.
Website scraping is coming soon. The endpoint is documented ahead of launch, and calls to it won’t work until it’s released. You can read the Scrape Website docs now.
Give your agent company websites
Website scraping is coming soon. When it launches, add scrape tokens to any paid Standard Subscription.