Website Business Evidence
Pricing
from $5.00 / 1,000 domain evidence cards
Website Business Evidence
Extract source-grounded business contacts, addresses, and organization social profiles from public websites, with one explicit evidence card per domain.
Pricing
from $5.00 / 1,000 domain evidence cards
Rating
0.0
(0)
Developer
Roman V
Maintained by CommunityActor stats
0
Bookmarked
5
Total users
1
Monthly active users
11 days ago
Last modified
Categories
Share
Turn a list of public business websites into explicit, source-grounded evidence cards. Each valid, distinct requested domain is eligible for one result, including domains with no supported contacts and domains that fail to load. Delivery stops at the run's charge limit; check emitted_records and charge_limit_reached in OUTPUT for incomplete delivery.
Why this Actor
- Every email, phone, address, and organization social profile includes its source URL and a bounded evidence excerpt.
- Phone extraction requires a
tel:link or an explicit phone label. Calendar dates, registration IDs, and loose digit sequences are not treated as phone numbers. - Primary contacts are selected deterministically. The result explains why a value was chosen or why no eligible value exists.
- Duplicate domains and canonical-equivalent contacts collapse before billing.
- Delivered cards have explicit empty, partial, or failed outcomes where applicable. A charge-limited run may stop before delivering all requested domains.
- The crawler is HTTP-first and bounded to likely public contact, about, legal, imprint, company, team, terms, and privacy pages.
- Gzip, deflate, and Brotli bodies are decoded before text parsing, with the configured limit enforced on both transfer and decoded bytes.
Input
Add one to 50 public HTTP or HTTPS websites to urls. Each entry is an object with url and an optional label.
The defaults visit up to eight pages per domain, include public addresses and organization social profiles, and process four domains concurrently. Lower maxPagesPerDomain for a faster home-page-only or shallow check.
Both transferred and decoded page bodies are limited to 1,000,000 bytes by default. Some modern sites exceed this. If a card reports RESPONSE_TOO_LARGE, explicitly set maxContentBytes up to 3,000,000 and consider lowering maxConcurrency. Larger bodies remain unsupported; no automatic retry or limit increase occurs. This failure does not mean that the site has no public contacts.
The default input uses Apify as a positive commercial example and Python.org as a public nonprofit example. Both are intended to return useful, source-grounded evidence rather than demonstrate a failure state.
Output
Each dataset row includes:
- A stable evidence card ID, requested URL, final URL, label, domain, and timestamp.
SUCCESS_WITH_EVIDENCE,SUCCESS_NO_EVIDENCE, partial, or failed status with a plain-language reason.- Source-grounded emails, strict phones, public addresses, and organization social profiles.
- A deterministic primary email and phone plus the selection reason.
- Every visited page and any page-level errors.
- Counts for pages and every evidence type.
The default key-value-store OUTPUT record contains run totals, invalid inputs, duplicate-domain counts, and maximum-charge status.
OUTPUT.status is PARTIAL when an input is invalid, a domain fails or returns a partial card, or delivery stops before all collected cards are emitted. A complete check with no supported contacts is not itself a failure. Reaching the charge limit on the last delivered card also does not make an otherwise complete result partial. This product-summary status is separate from the platform run's execution status.
The final log also reports failed, partial and no-evidence domain counts and page-error codes. These describe collected cards, including any not delivered because the charge limit stopped delivery; emitted_records separately counts delivery. Partial and no-evidence counts can overlap.
Malformed or unsupported compression, oversized decoded content, unsupported content types, binary-looking bodies, unsafe destinations, and cross-domain redirects produce explicit failed or page-level error outcomes. Raw runtime exception detail stays in logs and is not copied into status_reason.
Evidence and confidence
An email found in a visible mailto: link has stronger evidence than a loose text match. A phone found in a tel: link has stronger evidence than a labeled text value. Social links must be supported profile URLs, and non-LinkedIn profiles found away from the home or contact page must match the organization identity.
The Actor does not verify mailbox deliverability, phone ownership, or whether an address is currently staffed. It reports what the public website supports and shows the evidence used.
Safety and scope
This Actor reads public HTML and plain-text pages over HTTP or HTTPS. It rejects local, private, reserved, and non-public network destinations, including redirects. Every redirect and final URL must also remain within the registrable-domain boundary established by the requested URL; normal host and subdomain redirects within that boundary remain supported. It limits redirects, pages, request time, transfer and decoded response size, and concurrency. It does not log in, bypass access controls, submit forms, render JavaScript, or collect contacts from private sources.