Find companies and people
Discover prospects with structured filters and paginate without treating candidates as verified matches.
For a new saved list through MCP, use sheets_research: describe the companies or professional contacts, criteria, target count and typed fields. Leadbase researches the matches into a new private sheet. Follow progress with sheets_research_status, inspect results with sheets_query, and read enrichment_get for job outcomes and costs. Use sheets_enrich to qualify or fill details in existing rows. See research entry points.
For direct structured search through the HTTP API, use prospects.company_search for companies and prospects.people_search for people. These provider operations are not public MCP tools. In the app, open the matching Catalog entry and inspect its research form. For HTTP execution, read the operation descriptor first: supported filters and required null fields come from its live schema.
Search with explicit criteria
Company search can narrow a market using supported attributes such as geography, industry or company characteristics. People search targets supported person and employment attributes. Use exact identifiers when you already know the organization or person. Do not encode all criteria in a company-name field.
Read the descriptor's pagination fields and limits. Request the next page deliberately and use a different idempotency key for each distinct page. An example using a page size of one is a small test, not the service's universal maximum.
Separate discovery from qualification
A search record is a candidate. If your definition includes facts that search filters cannot establish, run enrichment on the candidates. For example, finding a company by location does not prove it currently recruits engineers. Ask that question separately, inspect evidence and retain the candidates that match.
Deduplicate by a stable identifier such as a domain or profile URL before enriching repeatedly. Start with a small page, measure its relevance and inspect actual credits before increasing volume.
Direct HTTP search returns when its page is ready; the research entry points return accepted work to inspect later. Preserve the original request and retry identity if a response is interrupted. See API recovery and Credits.