ravnA one-page account brief from a domain name.
Give Claude a domain and it can read the company's own site with scrape_website, pull size and industry with domain_to_company, and find the people in the role you sell to with search_people. You get a short brief you can check and act on.
Three tools, one brief
- / 01
Connect Ravn
Add the connector in Claude under Settings, Connectors, Add custom connector, using
https://api.askravn.com/mcp. See the MCP server page for Claude Code and Cursor. - / 02
Read the website
scrape_website returns the site title and description, address, public emails and phones, social links and detected technologies. It respects robots.txt.
- / 03
Add firmographics
domain_to_company runs on your own People Data Labs key (
ravn keys add pdl) and adds legal name, employee count, size band, industry, founded year, location, LinkedIn URL and NAICS codes. - / 04
Find the people
search_people with
company_domainand the role you sell to, also on your own People Data Labs key. Keep the limit small; two or three named people is enough for a brief. - / 05
Ask for the brief
Have Claude write what the company does, how big it is, what it runs on, who to talk to and one reason to reach out now. Ask it to cite which tool each fact came from.
The prompt and the raw calls
# In Claude Research acme.com with Ravn. 1. scrape_website for what they do, their tech and public contact details. 2. domain_to_company for size, industry, founded year and location. 3. search_people at company_domain acme.com for up to 3 people in sales operations or revenue operations, director level or above. Write a one-page brief: what they do, size, tech stack, who to contact and why now. Mark each fact with the tool it came from. Do not guess anything the tools did not return. # The same calls from the CLI npx ravn scrape_website --domain acme.com npx ravn domain_to_company --domain acme.com npx ravn search_people --company_domain acme.com --department "sales" --seniority director --limit 3
Per brief and for a batch
Every tool is one lookup. scrape_website is our own check; domain_to_company and search_people run on your own People Data Labs key.
| Step | One account | 25 accounts |
|---|---|---|
| scrape_website | 1 lookup | 25 lookups |
| domain_to_company | 1 lookup | 25 lookups |
| search_people (3 people) | 3 lookups | 75 lookups |
| Total | 5 lookups | 125 lookups |
Misses cost nothing, so a domain with no company record or fewer people costs less. All of it fits inside the free plan's 1,000 lookups; the people and company records are billed by your own provider. See pricing.
Keep the brief honest
Telling Claude to mark each fact with its source and not to fill gaps keeps the brief useful. If scrape_website finds no address, the brief should say so, not invent one. Ask for the last_updated_at date on each person so you can see which records are fresh.
The website scraper API, company enrichment API and people search API pages list every field each tool returns.
Before you reach out. Contact people only where you have a lawful basis, honour every opt-out, and never use this data for credit, employment, insurance or housing decisions (FCRA uses are forbidden by our terms). People who opt out at our opt-out page are never returned again. The full rules are on the compliance page.
Does scrape_website read every page of a site?
It reads the company's public site to pull out name, description, logo, address, emails, phones, social links and technologies. It respects robots.txt, so a site that blocks crawlers returns less.
Can I research a list of accounts at once?
Yes. Give Claude the list and ask it to confirm the lookup total first. For a file, run the CLI enrich command once per tool, for example with scrape_website and then domain_to_company.
Will Claude use LinkedIn for the people?
People lookups run on your own People Data Labs key, not on scraping LinkedIn. Records include a LinkedIn URL where the provider has one.
More in use cases
Every page uses the same account, plan and tools. See pricing →