Target Company URL Analysis: From One Domain to an Account Brief

Turn a company URL into an account brief in one prompt. Paste this into your agent:

Set up https://monid.ai/SKILL.md, and then use Monid to take https://vercel.com and tell me in three sentences what the company is: read its homepage description, detect the technologies the site runs, look up size, founding year and funding, and count the known contacts by department.

One prompt turns a URL into a brief. 1 min setup · 30-60 hrs saved.

Someone hands you a URL and nothing else: a prospect from a form, a competitor from a deck, a company on a list. You want to know what it is, how big it is, what it runs and whether anyone there is reachable. This guide shows the quickest way to get that from a bare domain, what each call returns (we ran every one on 2026-10-16 against vercel.com), and the traps we hit on the way. It runs through Monid, the OpenRouter for agent tools: one key, one balance, every tool priced per call.

What is the easiest way to analyse a company from its URL?

Monid. Install it once, add a key, and your agent fetches the page, detects the stack, resolves the company and counts its contacts with no vendor account for any of them.

• Pay per call. Each call bills on its own: $0.001 for the page, $0.0049 for the company profile and $0.012 for the live technology list (Verified 2026-10-16 on the company enrichment tools page). The contact count and the name-to-domain lookup are free. Buy calls as you need them: no monthly plan and no minimum top-up. Discovery and inspection cost nothing; only a run bills.
• One install, every source. The same key and balance reach the page fetcher, the firmographic index and the technology detector. Wiring those three yourself is the 30 to 60 hours this page claims to save.
• Lives inside your agent. Your agent reads SKILL.md, learns discover, inspect and run, and the tools are there whenever you paste a URL. Setup is about a minute.

A full brief on one company, page plus profile plus stack plus contact count, cost $0.0179 (Verified 2026-10-16).

How do you analyse a company URL step by step?

Four calls in a fixed order, the page first and the index second, so you notice when the index disagrees with the site.

For agents

Grab an API key at app.monid.ai, then paste this to your agent and hand it the key:

set up https://monid.ai/SKILL.md

It learns the whole discover, inspect, run workflow itself. More in the agent quickstart.

For humans

npm install -g @monid-ai/cli
monid keys add -k <your-api-key> -l main

Step 1. Read the page

monid run -p firecrawl -e /scrape -i '{"url": "https://vercel.com", "formats": ["markdown"]}'

Read data.metadata.ogDescription first and the markdown second. The metadata is the sentence the company wrote about itself; the body is navigation and banners.

Step 2. Read the server

monid run -p dataforseo -e /domain/domain-technologies -i '{"target": "vercel.com"}'

Pass a bare domain, without https:// or www. The answer is grouped by category, so you read servers, analytics, frameworks and CMS separately.

Step 3. Resolve the identity

monid run -p hunterio -e /companies/find --query '{"domain": "vercel.com"}'

Strip the scheme and any path first. This call wants a bare domain, and an unknown domain returns a 404 that costs nothing.

Step 4. Map the people

monid run -p hunterio -e /email-count --query '{"domain": "vercel.com"}'

Read the department object, not just the total. The shape of the organisation is in there.

Step 5, only if step 3 came back thin

A URL can be a campaign microsite, a regional subdomain or a redirect target. Resolve the brand name back to its main domain with the free hunterio/domain-finder and run steps 3 and 4 again on that:

monid run -p hunterio -e /domain-finder --query '{"company": "Vercel", "limit": 3}'

Or hand the whole job to an agent

Can you analyse a company URL yourself with Python?

Yes, and the first step is easy: requests plus an HTML parser gets you the title, the meta description and the Open Graph tags in a few lines. The rest is where the hours go.

Reading the stack means fingerprinting headers, scripts and markup against a rule set that someone has to keep current. Pages behind bot protection need a proxy or a headless browser. And headcount, funding, founding year and contact counts are not on the page at all: they come from an index you would have to license or build. A script covers the first call of this guide and none of the other three.

There is a rules side as well. What you may fetch and store depends on the site and on where you operate, so read the terms of the sites you analyse and the rules that apply to you.

That is the case for Monid for anything you will run more than once: the page fetch, the live detector and the company index are maintained for you, the output is JSON instead of HTML you parse, and a whole brief is $0.0179 (Verified 2026-10-16).

What does a company URL tell you, and what does each record contain?

A URL is three lookups wearing one string: the page, the server and the identity behind the domain. We ran all of them on 2026-10-16 against vercel.com, so the fields below are what came back.

The page

firecrawl/scrape returned HTTP 200 and a markdown body, plus a metadata object. The fields that matter:

{ "ogSiteName": "Vercel",
  "ogDescription": "The autonomous stack for every app and agent.",
  "title": "Agentic Infrastructure - Vercel",
  "robots": "index, max-image-preview:large",
  "og:url": "https://vercel.com",
  "statusCode": 200 }

The page is the only source that is authoritative about intent, and the only one that changes the day the company changes it. ogSiteName is the brand as the company spells it, robots says whether it wants the page indexed, and statusCode tells you whether you fetched the real page.

The server

dataforseo/domain/domain-technologies read the live site and returned the technologies grouped by category:

servers          cdn: Amazon S3 | web_servers: Next.js | paas: Vercel, Amazon Web Services
analytics        Linkedin Insight Tag
web_development  javascript_frameworks: Next.js, React | static_site_generator: Next.js
content          cms: Payload CMS

That is seven distinct technologies, plus the page title, description, a domain_rank of 586, the country and language, and the last_visited timestamp.

The identity

hunterio/companies/find returned the profile of the company that owns the domain:

name           Vercel            type            private
foundedYear    2015              location        94133, San Francisco, California, United States
employees      1K-5K             employeesCount  1026
trafficRank    very_high         raised          863000000
tech           46 entries        fundingRounds   6 (Series A 2020-04-21 to Series F 2025-10-30)

It also returned a description, tags, industry codes, social handles, a phone number and the addresses published on the site. The six funding rounds add up to the raised figure.

The people

hunterio/email-count is free and returned the contact map:

total 508   personal 460   generic 48
it 166   sales 65   support 35   operations 34   hr 24   marketing 23
finance & accounting 13   legal 12   executive 11   management 11   product 9
seniority: junior 3   senior 98   executive 81

What the four say together

A private San Francisco company founded in 2015 with roughly a thousand people, 863 million dollars raised across six rounds, running Next.js on its own platform, with 508 known contacts of whom 166 sit in IT. That is an account brief, from a URL, and nothing in it needed the company's name.

What you can build with it

• An account research card. One prompt per prospect, with the positioning sentence, size band, stack and department mix on one screen.
• A reachability check before you spend. The free count tells you whether a domain has anyone worth contacting before you pay for names; Hunter.io domain search vs email finder shows the paid step that comes after it.
• A qualification rule. Combine size, stack and department mix into a number across a list, which is the subject of the lead scoring guide.

What goes wrong when you analyse company URLs?

Five things hide until you run more than one company. All five showed up in the runs of 2026-10-16.

Why is the positioning sentence not in the page body?

Because the homepage body is navigation. The markdown for vercel.com opened with a skip-to-content link and a "Read the recap" banner before reaching a hero line, and the company's own one-sentence description sat in the metadata. ogDescription is one sentence, maintained because it is what appears when anyone shares the link.

Read the metadata first and treat the body as supporting text. Extracting the positioning from the body means writing heuristics about which heading is the real one.

Why did the first scrape come from a cache?

metadata.cacheState came back as hit with a cachedAt stamp hours earlier. The endpoint reuses a recent copy of the page unless you ask otherwise, which is fast and cheap but not live.

Set maxAge to 0 when freshness matters, for example when you are checking what a company says today after an announcement:

monid run -p firecrawl -e /scrape -i '{"url": "https://vercel.com", "formats": ["markdown"], "maxAge": 0}'

A page that answers 403 or 404 is still billed, so read metadata.statusCode before you retry a blocked URL.

Why do the live detector and the index list different stacks?

They read different things. The live detector found seven technologies on the site. The company profile listed 46 tech entries, including Netlify, Nuxt and Azure alongside Next.js and Vercel, which is an accumulated list over time, not a reading of today's page.

Use the live read for what the site runs now and the index for what the company has been seen using. Do not merge them into one list without keeping the source on each row. The same split is measured in the technographic data guide.

Why do the numbers differ between runs?

An index moves, and its records have an age. The profile carried an indexedAt date of 2026-06-18, months before our run, so its headcount is a reading from that date, and the email count is a separate index with its own refresh. Our run put employeesCount at 1026 and the band at "1K-5K".

Store the date next to every figure and read headcount as a band, not as an exact number.

Why does a URL sometimes describe the wrong company?

Because the calls answer about the domain you give them, correctly. A campaign microsite, a regional subdomain or a product domain owned by a parent returns thin or misleading results. The free name-to-domain lookup for "Vercel" returned vercel.com with 508 known addresses and vercel.me and vercel.lol with none, so the email count alone separates the real domain from lookalikes.

When the profile comes back thin, resolve the brand to its main domain and rerun. Best API to search a company's homepage from its name covers that direction.

Which endpoint should I use for which job?

Every row was inspected on 2026-10-16 and the latency is the endpoint's typical (median) run time. Billing shape is given here, and the current figure is on monid.ai/tools.

The last two rows are free. Start with them whenever you only need to decide whether a domain is worth the three paid calls.

When is another route a better fit?

The URL is not the company. Every call answers about the domain you gave it. For a microsite or a regional subdomain, resolve the main domain first.

You need the legal entity. Registered name, officers and filing status come from a registry, not from a homepage or a firmographic index. That is the business entity search.

The site is rendered on the client or blocked. A fetch returns what the server sends. A page that builds itself in the browser, or refuses the request, gives the page call little to read, and the free count is the only signal left.

You are working a very large list. A million domains is a licensed dataset job, not a per-call job; B2B data providers, dataset or API lays out where the line sits.

You are republishing the data. Using headcount and stack internally to qualify accounts is ordinary. Publishing someone's contact counts or profile as your own product is a different act, and the rules differ by country, so check yours.

And the disclosure: this is Monid's blog and we resell all five endpoints. The most useful move here costs nothing on any vendor: read the Open Graph description before anything else.

Conclusion

The quickest way to analyse a company from its URL is four calls through Monid: the page for $0.001, the live stack for $0.012, the company profile for $0.0049 and the contact count for nothing (Verified 2026-10-16), one key for all of them, about a minute to set up and ready inside your agent whenever someone hands you a link. A brief on one company was $0.0179.

Two things matter more than the choice of tool. Read the metadata before the body, because the company's own positioning sentence is one maintained field while the body is navigation. And keep the source and the date on every figure, because the live detector and the index list different stacks, and an index record carries the date it was built.

How we estimated the time saved: building your own URL-to-brief pipeline is roughly 10 to 20 hours each for a page fetcher that survives blocking, a stack detector, and a company and contact index. That is 30 to 60 hours. The 1 minute is the install, key and first call once you already have a Monid key.

Free next step: run monid inspect -p hunterio -e /email-count, then call it for a company you know and read the department object before anything else. Start at monid.ai.

FAQ

Published September 2026.