Research companies, not people
Sales and partnership teams spend a lot of time qualifying accounts: is this company in our segment, does it sell the kind of product we integrate with, is it growing, is it publishing about the problem we solve? Those are questions about companies, and company websites answer them in public.
What Siftzo does not support is harvesting personal data such as names, email addresses and phone numbers of employees for outreach lists. Under data protection rules like the GDPR, personal data needs a lawful basis, a purpose and limits on what you collect, and bulk-collected contact details rarely meet that bar. Siftzo's acceptable use policy prohibits it. This page is general information, not legal advice; ask your own adviser about your case.
Step 1: build the account list
Lead research starts with a list of company websites. You may already have one from your CRM, an event or an industry directory you are allowed to use. If not, the SERP API can help you find companies in a segment: search for the terms a company in that segment would use to describe itself, per market, and collect the domains from the organic results.
Clean the list before you spend requests on it. Normalise each entry to the home page URL, remove duplicates and directories, and drop domains that are clearly not companies in your segment. A short, well-defined list gives better results than a long one, and it keeps the number of requests, and the load on other people's websites, in proportion.
Step 2: read the basics from page metadata
Start with each company's home page and the Data API metadata source. It returns the page title, description, canonical URL, language, site name, Open Graph tags, linked feeds and every JSON-LD block, which often includes the organisation's name, logo, description and links to its other pages.
curl -G https://api.siftzo.com/v1/data/metadata \
-H "X-API-Key: $SIFTZO_API_KEY" \
--data-urlencode "url=https://www.example.com/"The metadata record is contact-free by design. Siftzo removes email, telephone, faxNumber, address and contactPoint fields and whole Person objects from JSON-LD, and never collects mailto: or tel: links. You get what the company says about itself, not who works there.
Step 3: qualify with a few targeted pages
Metadata tells you what a company is; a few more pages tell you whether it fits. Typical pages to read are products or services, industries or customers, pricing, integrations and the news or blog section.
- Fetch them with the Scraper API and
format: "markdown"to read clean main content. - Or use a small crawl with
includeglobs like/products/**and a lowmax_pages, so you only touch what you need. - Use
/v1/data/articleon news pages to get titles and publication dates and see what the company has been announcing.
Step 4: turn findings into fields
To compare hundreds of accounts, you need the same fields for each. Define a JSON Schema with company-level facts only, for example industry, products, target_customers, offers_free_trial and mentions_topic, and send it with the page to AI extraction. Every result matches the schema, so you can filter and score accounts in your CRM or a spreadsheet.
Leave personal fields out of the schema, and say so in the prompt. If your team needs a contact at a company, use the channels the company itself offers, such as its contact form or sales page, rather than a scraped list.
Keep the source URL and date with every field so a salesperson can check the claim before using it. Team API keys come with their own logs and credit caps, which makes it easy to give a research project a fixed budget. See pricing for credits per request type.
Doing it responsibly
- Company-level facts only; no personal data of employees or customers.
- Public pages only, and the site's terms respected.
- A clear purpose and a retention period for what you store.
- Low request volume: a handful of pages per company is usually enough.
Siftzo is hosted in the EU. Read the responsible use guidelines and the acceptable use policy before you start.
Try it on your own data
1,000 free credits every month, no card needed. Failed requests are never billed.
Questions
Can Siftzo find email addresses or phone numbers of employees?
No. The Data API has no sources for people, the metadata record strips contact fields and Person objects from JSON-LD, and mailto: and tel: links are never collected. Harvesting personal data is against the acceptable use policy.
Is it legal to research companies from their websites?
Reading public, company-level information is common practice, but the rules depend on what you collect and how you use it. This is general information, not legal advice; check the site's terms and ask your own adviser.
How many pages should I fetch per company?
Usually only a handful: the home page, products or services, and news. A small crawl with include globs and a low max_pages keeps both cost and load low.