People Search

Find people — a single named person, or a whole list that fits a profile — with the people_search tool. It runs a LinkedIn profile search and returns structured profiles: name, title, current company, location, and LinkedIn URL. It works out of the box (Kortix-proxied, billed to the account) or with your own APIFY_TOKEN.

Using the tool

people_search(query="climate tech founder", titles="Founder,CEO", locations="Lagos", num_results=10)
  • query — free-text: name, role, focus, or keywords. This is the main lever.
  • titles — comma-separated current job titles to match (e.g. "Head of Talent,Recruiting Lead").
  • locations — comma-separated city / region / country (e.g. "Singapore").
  • num_results — 1–25 (default 10).
Each result is { name, title, company, location, linkedin_url, email }. The search is a real run, so it can take a little while — let it finish.

Get the query right

The result is mostly as good as the query + filters:
  • Spell out and abbreviate the role when it helps — "developer relations" vs "DevRel"; "Head of Talent" vs "recruiting lead".
  • Anchor with what you know — company, city, school, seniority, a distinctive skill. More anchors → cleaner results.
  • One person? Put the name in query (add company/location if you have them) and keep num_results low.
  • A list? Lead with titles + locations and a focused query; raise num_results.

Refine the results

people_search returns people who match — tighten before you deliver:
  1. Filter to true matches. Judge each person against the real criteria, phrased as exclusions ("drop anyone NOT currently a Head of Talent, NOT at a fintech, NOT in Singapore"). Note in one line why each kept person qualifies.
  2. Enrich when you need depth. The profiles are concise (title / company / location). For bio, work history, education, or projects, scrape_webpage the person's linkedin_url or another public page (personal site, company team page, conference bio, Crunchbase). LinkedIn itself is often gated — corroborate from a second public source when it is.

Output format

For a named lookup or dossier:
markdown
# [Full Name] — [Current Title], [Company]

**Profile:** [LinkedIn link]

## Snapshot
[2–3 sentences: who they are, what they're known for, current focus.]

## Background
| Field | Detail |
|-------|--------|
| **Current role** | [Title at Company] |
| **Location** | [City / region] |
| **Previously** | [Notable prior roles] |
| **Education** | [Institutions] |
| **Focus / skills** | [Areas, specialties] |

## Sources
- [Source 1](URL)
For a candidate list, link each name to their profile, one row per match:
markdown
# [Search description] — [N] matches

| Person | Role & Company | Location | Why they match |
|--------|----------------|----------|----------------|
| [Amara Okafor](https://…) | Founder, GreenGrid (climate tech) | Lagos | Series-seed climate startup; ex-Tesla energy |
| [Wei Chen](https://…) | Head of Talent, Paywave | Singapore | Leads TA at a fintech, 3 yrs in role |

Richer Lookup: When a People-Data Connector Is Connected

The web path above (people_search) works standalone and covers most asks. If the project also has a people-data connector wired up — an enrichment, prospecting, or contact-data provider such as Apollo, People Data Labs, Clearbit, Clay, RocketReach, Hunter, Lusha, or ZoomInfo — prefer it when precision and verified fields matter, or for company-level lookups the web tool doesn't cover.
Check what's connected:
bash
kortix executor connectors          # or the `connectors` MCP tool
A connector qualifies if its job is people/company data. (CRMs like Salesforce/HubSpot can enrich known contacts too, but don't discover new ones — treat them as a supplement.) Nothing connected → stay on the web path above; it already answers most asks, and it's fine to offer to wire one up without blocking the request.
Two operations cover almost everything once a connector is live — exact action names vary per provider, so inspect its tool schema and pick the closest match:
  1. Find / shortlist — the connector's people-search or prospect-search action takes structured filters (title, company, location, seniority, headcount, industry) and returns candidates from a curated index. Fan out the same way as the web path: vary title wording and scope across a few calls, then merge by a stable key (profile URL or provider person-id). Coverage is cleaner than a web sweep but not exhaustive — a miss is not proof the person doesn't exist; fall back to the web path for the gaps.
  2. Enrich — the connector's enrich / profile-lookup action resolves a known identifier (name+company, email, domain, or profile URL) into a fielded record: verified email, phone, employment history, education, org chart. This is the high-value path for turning a list into clean rows with no scraping. Batch where the action supports it; respect per-call ceilings and credit budget.
Company lookup. For account-level questions — firmographics, headcount, org chart, tech stack, funding — most people-data connectors expose a company/organization-search or company-enrich action alongside the people actions. Use it the same way: search by domain or name, enrich into a fielded company record. Without a connector, get company facts from the web path (web_search + scrape_webpage on the company site, LinkedIn company page, Crunchbase) or hand off to account-research for a full company dossier.
Filter — same discipline as the web path. A curated index still returns loose matches. Phrase criteria as exclusions and judge each candidate against the fielded data returned; don't skip this step just because the connector felt more authoritative.
Wiring up a connector (when none exists). Don't send the user to a dashboard — use the credentials flow (kortix-system → Credentials & setup links):
  • Pipedream-backed app (most prospecting tools): kortix executor add apollo --provider pipedream --app apollo, then kortix executor connect apollo to mint a 1-click link.
  • API-key provider (e.g. a People Data Labs key): kortix secrets request PDL_API_KEY --scope connector.
Surface the URL, end your turn, and verify it landed (kortix executor connectors / kortix secrets ls) when the user returns.
Output differences on the connector path:
  • Source line names the connector, not "Web Search" — **Found via:** Apollo.
  • Verified-contact fields carry through — label an email that came from the connector email_verified: true. Anything not returned by the connector stays absent or email_verified: false. Never fabricate or pattern-guess, on either path.
  • Respect compliance gating — if an action returns a permission/region error, surface it plainly rather than working around it.

Gotchas

  • The name is the link. Hyperlink each person's linkedin_url on their name — [Amara Okafor](https://…) — Founder, GreenGrid. Never a bare name or a trailing raw URL. For CSV (no markdown), keep a profile_url column.
  • Never invent emails. If email is null, leave it out or mark email_verified: false. Never pattern-guess first.last@company.com.
  • Corroborate moves. Titles and companies change. For a high-stakes claim, confirm the current role from the profile or a second source.
  • Gated profiles. LinkedIn pages often block scraping — lean on the structured fields people_search already returned and corroborate elsewhere instead of fighting the login wall.
  • Results count ≠ matches. num_results=25 returns up to 25 candidates, not 25 perfect fits — the filter step produces the real number.

Tips

  1. Give anchors — company, city, school, "spoke at X" — they cut noise sharply.
  2. Name the real criteria so the filter step can do its job.
  3. State the goal — a recruiting shortlist, an outreach list, and a single dossier need different depth.

Related skills

  • account-research — full picture of a company or person before outreach.
  • deep-research — multi-source, fact-checked report when the question is broader than "who."
people-search — Kortix Marketplace | Kortix