Live numbers, updated every hour

Data quality

What we have, how fresh it is, how each detection is made, and what we don't have yet. Every number comes from the same tables the app and the API read.

Ady, our crawler, is running

0 companies checked in the last 24 hours. The latest real companies, as they are checked:

About the crawler →

34M

websites profiled

GET /api/v1/lookup?domain=

196M

technology detections

GET /api/v1/tech/{slug}

4,281

technologies tracked

GET /api/v1/tech

0

profiled in the last 24 hours

GET /api/v1/lookup?domain=

0

new customers detected, 30 days

GET /api/v1/signals?days=30

How we detect

Every detection records how we found it. Most come from the code on a company's homepage; about one in ten comes from DNS records, which show software the website never mentions (email, security, help desks, sign-in providers).

Where detections come from

A random sample of 39,515 active detections, 27 October 2026.

Website codescripts, page globals, signatures88.7%
DNS recordsMX, TXT, SPF9.7%
HTTP headers1.7%

By method

The label you see next to each company in a list.

page code72.4%
MX record4.8%
script4.7%
page global4.6%
DNS TXT record3.9%
meta tag3.6%
signature2.4%
HTTP header1.7%
SPF record1.1%
iframe0.9%
preconnect0.1%
cookie0.1%

How complete a profile is

Measured on 342,036 profiles on 29 October 2026. Re-measured every week.

Company data

Share of profiled companies with each field.

Company name99.8%
At least one technology85.0%
Industry16.5%
Email format12.3%
Country6.8%

Contacts

Secondary to the technology data, and shown with their status. Unverified addresses are never presented as verified.

At least one email address29.3%
A verified email address0.9%

24M

Found

805K

Verified

2.4M

Waiting to verify

227K

Dead, hidden

Freshness

How long since each profile was last checked. Every detection also carries the date we last saw it.

Last checked

34,269,550 profiles, 29 October 2026.

In the last 7 days6.7%
7 to 30 days ago73.2%
30 to 90 days ago2.1%
More than 90 days ago14.1%
No check date recorded3.9%

Tools started, per day

Companies we saw start using a tool. History begins September 2026.

16 OctToday, so far

What we don't have yet

  • Stops

    We record when a company starts using a tool. Stopping is harder to prove (a page can fail to load), so a tool will only count as stopped after two clean re-checks. That check is not live yet, so we do not report stops.

  • Accuracy per product

    Not published yet. We will hand-check samples for the products we sell lists of, and publish the results here, product by product.

  • Software that leaves no public trace

    We see what a company's website and DNS records show. Internal tools, and apps that only load after sign-in, are invisible to us.

  • Industry and country for most companies

    Industry is known for 17% of profiles and country for 7%. Both are filled in only where a source shows it.

  • Revenue and company size

    Revenue is not collected. Our team-size estimate from websites is not reliable enough to show yet, so lists have no size column; where a company states its headcount we show it, labelled as an estimate.

How it is collected

Our crawler, AgentDataBot, names itself on every request and obeys robots.txt. It reads a company's public pages and response headers, and we read the public DNS records of the domain.

We take nothing from LinkedIn and do not scrape Google. Any company or person can ask us to remove their data on the data removal page.

What a company record contains

Field by field, with where each one comes from. A field is empty when we did not find it; nothing is guessed and shown as fact.

TechnologiesEvery tool we detected, with its category, how we found it (the script, DNS record or header) and when we first and last saw it.
ChangesTools the company started using, with the date we saw each one appear.
CompanyName, domain, sector and vertical (classified from the company’s own description), business model, B2B or B2C, country where known, and headcount where the company states it (labelled as an estimate).
Web signalsWhether it has a pricing page, a free trial, API docs, a blog or a careers page, and its main call to action.
ContactsThe email format, published addresses with their verification status, and named people with job titles where the company lists them. LinkedIn URLs only when the company links them.
ProvenanceWhen the company was last checked, how complete its profile is, and the outcome of the last crawl.

The full response is in the API docs.

What is next

In the order we are doing it.

  1. 01Confirm stops: a tool will count as stopped after two clean re-checks without it. Then stops appear here, in lists and in alerts.
  2. 02Accuracy per product: hand-checked samples for the products we sell lists of, published on this page.
  3. 03A monthly re-check of every company in our customer lists, so their “last checked” stays under 30 days.
  4. 04Country and industry for more companies, from contact pages, DNS records and the companies’ own descriptions.

Something you need that is not here? Tell us.

Query any of it

Single look-ups through the website and the REST API work without an account (300 requests every 5 hours). A paid plan adds full lists, CSV downloads and alerts, and the same data in the API and the MCP server. Pro and Business both start with a 7-day free trial, no card needed.

# no key
curl agentdata.run/api/v1/tech/intercom
curl "agentdata.run/api/v1/signals?days=7"
curl "agentdata.run/api/v1/companies?tech=intercom"
# in Claude Code
claude mcp add --transport http \
  agentdata https://mcp.useagentdata.com/mcp
# opens a page to connect your
# AgentData account

# then ask
"Which companies started using
 Intercom this month?"