Contact Evidence Enricher
Pricing
from $3.00 / 1,000 site checks
Contact Evidence Enricher
Enrich your company list with public contact details from each website: emails, phones, social profiles and addresses, each with its source page. Every row comes back unchanged; misses get a plain reason. $0.003 per site.
Pricing
from $3.00 / 1,000 site checks
Rating
0.0
(0)
Developer
ForLab Digital
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
15 hours ago
Last modified
Categories
Features
Share
Send a list of companies with their websites. Get every row back, unchanged and in the same order, with the public contact details found on each site: emails, phone numbers, social profiles (LinkedIn, Instagram, Facebook, X, YouTube, TikTok) and postal addresses. Every contact comes with the page it was found on, so you can check it.
- Your data stays intact. Your columns are kept as they are; new fields go under
enrichment. - Evidence, not guesses. Each email or phone has a
source_urland themethodit was read with. No guessed addresses, no pattern-generated emails. - A reason for every miss. A site with no result gets a plain-language
note: bot protection, robots.txt, a redirect to another domain, or no contact on the pages read. - Pay per site checked. $0.003 per website, with a hard cost limit per run. An optional paid Google search finds emails for sites that block crawlers.
Input
Supply exactly one of records (a JSON array of objects) or csvText (CSV with a header row):
{"records": [{"lead_id": "A-001", "company": "Example Ltd", "website": "https://example.com"}],"websiteField": "website","requiredContact": "any","maxPagesPerSite": 3,"concurrency": 2,"searchFallback": false}
websiteField: the column that holds the website. Values without a scheme use HTTPS.requiredContact: which contact makes a site count as found.any(an email, a phone or a social profile; the default),either(an email or a phone),emailorphone. Every contact type found is returned in all modes.maxPagesPerSite: 1–5 page attempts per site (the home page plus likely contact pages). Robots.txt requests and redirect hops are extra requests.concurrency: 1–4 sites crawled at once.searchFallback: off by default; see "Paid search fallback".- Limits: 1–200 records and 2 MiB input per run.
Output
Each input record produces one dataset item in the original order. The original JSON values are kept under input, the zero-based position is row_index, and new fields are under enrichment:
{"row_index": 0,"input": {"lead_id": "A-001", "company": "Example Ltd", "website": "https://example.com"},"enrichment": {"site": "https://example.com/","status": "contact_evidence_found","requested_contact_evidence_found": true,"emails": [{"value": "info@example.com", "source_url": "https://example.com/contact", "method": "mailto", "domain_matches_site": true}],"phones": [{"value": "+441234567890", "source_url": "https://example.com/contact", "method": "tel"}],"socials": [{"platform": "linkedin", "url": "https://www.linkedin.com/company/example", "source_url": "https://example.com/", "method": "link"}],"addresses": [],"pages": ["https://example.com/", "https://example.com/contact"],"errors": [],"ownership_verified": false,"mailbox_verified": false,"charged": true}}
The values above are illustrative. The run also writes a summary to the key-value store OUTPUT record. Use the JSON output when you need the exact original values; Apify's CSV and Excel exports flatten nested fields.
What each field means
emailsandphones:value,source_urlandmethod(mailto,tel,visible_textorjson_ld). A contact found on several pages appears once per page, each with its ownsource_url. Phones come only fromtel:links and schema.orgtelephone; no country code is inferred.domain_matches_siteon each email: true when its domain equals the site's domain (ignoringwww.) or is a subdomain of it. Use it as a filter; parent-company, registered e-mail (KEP) and agency addresses come back as false.socials: profile links found in page links or JSON-LDsameAs, each withplatform,url,source_urlandmethod. Share buttons and single posts are dropped.addresses: postal addresses from JSON-LDaddress(withpartsand, when given,geo) or<address>elements. Free text is never parsed as an address.statusandnote: why a site gave no result.no_requested_contact,robots_denied,robots_unavailable,fetch_failed,redirected_to_other_site(the new address is inerrors[].redirect_to),crawl_delay_exceeds_limit,deadline_exceeded,internal_error,charge_limit_reached, andmissing_websiteorinvalid_urlfor a row without a usable website. When the site answered 401, 403 or 429, the note says it blocks automated visitors.charged: whether this row's site was billed.
A source URL means the text appeared on that page. It does not prove that the contact belongs to the company, that the mailbox works, or that the company agreed to be contacted. Check the evidence before outreach.
Run summary (OUTPUT record)
status_counts per site status, sites_blocked (sites that answered 401, 403 or 429), search_fallback (off, done with searched_sites and sites_found, skipped, or failed), charged_events and billing_mode (pay_per_event or off).
Pricing (pay per event)
site-checked, $0.003: each website in your list that was visited, whether or not contact details were found. Blocked sites count: they were tried. Sites cut short by the run deadline or an error, rows without a usable website, and sites past your cost limit are not charged.google-search, $0.008: each site searched on Google, only withsearchFallback: trueand only on paid Apify plans. On the free plan the search runs on your own account instead, and apify/google-search-scraper bills you directly (about $0.0045 per site searched).- The run never charges more than its maximum cost (
maxTotalChargeUsd, $2.50 by default). Sites past it are not crawled and come back with statuscharge_limit_reachedand a note; raise the limit and send them again. 200 sites cost $0.60, plus up to $1.60 if all are searched. - Charges are sent after the results are saved.
summary.charged_eventsand each row'schargedlist what is billed; a charge request that fails is logged ascharge_failedand is not billed.
Paid search fallback (searchFallback, off by default)
The crawl reads each site itself. Some sites block it or show none of the requested contact. Set "searchFallback": true to search those sites once on Google as "@domain" through apify/google-search-scraper.
- Cost: $0.008 per site searched on paid plans (see Pricing), about $0.0045 billed by apify/google-search-scraper on the free plan; charged even when nothing is found. Sites where the crawl found the requested contact are never searched; the search finds emails only. Phone-only runs (
requiredContact: "phone") never search. - Result: emails on the site's domain seen in result titles or snippets are added with
method=search_snippet, the result page assource_url, andsource_on_site. They come from third-party pages (social profiles, app stores, registries, review sites) and may be stale or wrong. - Yield varies: in a 61-site test it found an email for 22 sites, and a repeat search can return different results.
- Time: the crawl stops up to 5 minutes (never more than half the run) before the run deadline to leave the search its time.
- To keep only emails seen on the site itself, drop those with
source_on_site: false.
Use it from code or an AI agent
Start a run through the Apify API and read its dataset; no knowledge of this Actor's code is needed:
curl -X POST "https://api.apify.com/v2/acts/forlab~contact-evidence-enricher/run-sync-get-dataset-items?token=$APIFY_TOKEN" \-H "Content-Type: application/json" \-d '{"records":[{"website":"https://example.com"}]}'
The input schema, the dataset rows, the OUTPUT summary and the status codes above are the stable contract.
How the crawl behaves
- It follows robots.txt and its crawl delays up to five seconds. A robots.txt 404 permits crawling; an unreadable robots.txt stops that site.
- It reads only the submitted site. Redirects within the same host (including
wwwand HTTP to HTTPS) are followed; a redirect to another domain or subdomain stops the site withredirected_to_other_site. - Rows sharing a website are crawled once and the result is copied to each row.
wwwand bare hosts are separate sites and are charged separately. - No JavaScript rendering, proxies, CAPTCHA solving, logins, mailbox verification or outreach. Some sites will therefore give no result; the
notesays why. - Only public HTTP/HTTPS addresses on ports 80/443 are fetched. Private, loopback and cloud metadata addresses are rejected, TLS certificates are checked, and responses are capped at 2 MiB with at most three redirects.
- A run stops starting new requests 30 seconds before its timeout, so every row is still written; unfinished sites get
deadline_exceeded.