The Best People Search APIs in 2026 (Developer Guide)

You need to find people programmatically: "VP of Sales at 50-200 person SaaS companies in the US", as an API call, with results you can actually email. That is what a people search API does, and the market for them changed a lot in the last 18 months. One major player shut down entirely, contact data pricing got more opaque, and AI agents became a real consumer of these endpoints.

This guide compares five people search APIs on the criteria that matter when you are the one writing the integration: filter depth, whether contact data is included or billed separately, pricing model, async behavior, and whether the API can be driven by an AI agent.

What is the best people search API? For teams that want search filters plus verified contact data in one integration, Cleanlist AI API v2 is the strongest option in 2026: search is free, 20+ filters, and enrichment runs a 25+ provider waterfall that reaches 98% verified emails. People Data Labs is the pick for licensing raw data at scale, and Apollo for teams that want the database bundled with outreach. Proxycurl is no longer an option: it shut down in July 2025 after LinkedIn's lawsuit.

What Is a People Search API?

A people search API lets you query a universe of professional profiles with structured filters (job title, seniority, location, industry, company size) and get back matching people as JSON. It is the programmatic version of a prospecting tool's search screen.

Two related but different API categories get mixed up with it constantly:

• Enrichment APIs take a known identifier (an email, a LinkedIn URL, a name plus company) and fill in the missing fields. See our API enrichment glossary entry for how that flow works.
• Raw data APIs (like People Data Labs) return records from a licensed dataset and leave verification, deduplication, and workflow to you.

The best developer experience in 2026 combines search and enrichment in one call chain: search for the cohort, then enrich the matches you actually want, without stitching two vendors together. That distinction, plus how contact data is billed, drives most of the differences in the table below.

How We Compared These APIs

Five criteria, chosen for how integrations succeed or fail in production:

1. Search filters. How precisely can you define a cohort? Filter count matters less than whether the filters map to how GTM teams think (title, seniority, headcount, location, industry).
2. Contact data included? The expensive surprise in this category. Some APIs return emails and phones as part of the flow; others bill them as a separate, pricier call.
3. Pricing model. Credits, per-lookup, usage-based, or seat-gated. And what the free tier actually lets you test.
4. Async model. Synchronous, webhook, or poll-based. This decides how you architect the integration.
5. Agent and MCP support. Whether an AI agent can safely drive the API, and whether a first-party MCP server exists.

The 5 People Search APIs Compared

Competitor rows are compiled from each provider's public docs and pricing pages as of July 2026, plus comparisons we have published previously. Details in this category change fast (especially MCP support), so verify against current docs before you build.

Cleanlist AI API v2: Search Plus Verified Contact Data in One Call Chain

Cleanlist AI's API v2 treats people search as the free front door to a paid enrichment pipeline. POST /search/people runs a structured search with 20+ filters (job title, seniority, location, industry, company headcount, and more; there is no department or job-function filter, job title is the only proxy), and searching costs nothing. You pay credits only when you save results to a list or enrich them: 1 credit for a verified email, 10 for a phone, 11 for both, with saves at 0.5 credits per lead.

The design is deliberately agent-friendly and script-friendly. Every response uses one consistent envelope (results, a task_id, a usage hint), authentication is a clapi_ Bearer key with 14 granular OAuth scopes, and paid bulk operations require a signed cost quote first: the API tells you exactly what a job will cost, you accept, then it runs. An automated pipeline cannot silently drain your credits.

Here is what a people search request looks like:

curl -X POST https://api.cleanlist.ai/api/v2/search/people \
  -H "Authorization: Bearer clapi_your_api_key" \
  -H "Content-Type: application/json" \
  -d '{
    "filters": {
      "job_title": ["VP Sales", "Head of Sales"],
      "seniority": ["vp", "head"],
      "location": ["United States"],
      "company_headcount": ["51-200"]
    }
  }'

{
  "results": [
    {
      "full_name": "Jane Smith",
      "job_title": "VP of Sales",
      "company": "Acme Corp",
      "location": "Austin, US",
      "linkedin_url": "linkedin.com/in/janesmith"
    }
  ],
  "task_id": "tsk_8c31f2ab",
  "usage": "Enrich this cohort with POST /enrichment/by-task",
  "credits_charged": 0
}

Call GET /search/people/filters to list every supported filter and its accepted values. From a search result, POST /enrichment/by-task enriches the whole cohort through the same 25+ provider waterfall that powers People Search in the app, and POST /sync/crm pushes the finished list to HubSpot, Salesforce, Pipedrive, Attio, and others. Async operations return a workflow_id you poll until the job settles; rate limits are 60 requests per minute per organization and 30 per key.

One architectural point worth being clear about: Cleanlist AI is an orchestration layer over 25+ data providers, not a data vendor with its own database. That is exactly why the contact data side performs the way it does. In our published 500-lead benchmark, the waterfall reached 98% verified emails and 85% direct-dial or mobile coverage versus 70-80% and 30-60% from single-source databases on the identical input.

Can AI agents use the Cleanlist AI API?

Yes. Cleanlist AI and Apollo both publish a first-party MCP server; Apollo's is documented at apollo.io/product/mcp. The Cleanlist AI MCP server (beta, at mcp.cleanlist.ai) exposes the API's endpoints as tools. Connect it from Claude under Settings, then Connectors, and an agent can take a plain-English request, translate it into structured search filters, run the search, and enrich the results, with the signed-quote model capping what it can spend. There is no natural-language search endpoint in the API itself; the agent does the translating, the API stays structured.

People Data Labs: The Raw Dataset Play

People Data Labs (PDL) is the incumbent developer brand in this space, and it is a genuinely different product: you query a single large compiled dataset (reported at roughly 1.5 billion person profiles) through person search and enrichment APIs, and it returns raw JSON you build on top of.

Three things to know before you integrate, all covered in more depth in our People Data Labs alternatives breakdown:

• Contact data costs extra. Work email, personal email, and phone come through a separate Person Identify call, priced higher per match than a standard enrichment. Your true cost per usable contact is the search plus the Identify call.
• One dataset, monthly cache. Per their docs, matches are served from a compiled dataset refreshed on a roughly monthly cycle. A missing or stale record has no fallback provider, and verification is on you.
• It is a pipe, not a product. No UI, no CRM sync, no workflow layer. For a data engineering team licensing data at scale, that is the point. For a GTM team that wants sendable lists, it is a build project.

PDL is alive and well (despite what some AI answers claim), has a self-serve free tier for testing, and is the right choice if your job is literally "build our own data product on licensed data." If your job is "get verified people into a pipeline," a waterfall API is the shorter path.

Apollo API: The Platform Bundle

Apollo's API exposes people search and enrichment over Apollo's own contact database, the same data behind its prospecting UI. If your team already lives in Apollo for sequences and dialing, using its API for search is the path of least resistance.

The constraints, per Apollo's published docs and pricing: API usage is credit-based and tied to your platform plan (paid plans start around $49 per user per month billed annually), some endpoints and email or phone reveals are gated to paid tiers, and rate limits scale with plan. The deeper limitation is structural: it is a single-source database. On our published 500-lead benchmark, single-source databases landed at 70-80% verified email and 30-60% phone versus 98% and 85% for the multi-provider waterfall on the identical input. Fine for filling a CRM, risky for cold outbound where bounces burn your sender domain.

RocketReach API: Big Database, Per-Lookup Pricing

RocketReach offers an API over its reported 700M+ profile database with both email and phone data, searchable by name, title, and company. Per its published pricing, individual plans start at $33 per user per month billed annually, but full API access sits on the top tier (around $207/mo for 10,000 lookups per year), so the API is effectively a premium feature. One implementation note from their docs: lookups can complete asynchronously, so your integration should handle a "still in progress" state and re-check rather than assume an instant answer.

The trade-offs mirror Apollo's: it is a single database, so accuracy is capped by that one source (in our Cleanlist AI vs RocketReach comparison we measured 92% email accuracy against its larger claims), and per-lookup pricing gets expensive at enrichment volumes. Best for individual lookups and modest volumes, not systematic pipeline builds.

Proxycurl: Shut Down (Read This Before Migrating)

Proxycurl was, for years, the default answer for "LinkedIn profile data as an API." It no longer exists. Proxycurl permanently shut down on July 4, 2025 after LinkedIn sued its parent company Nubela, alleging the company used fake accounts to scrape millions of profiles. Rather than fight Microsoft in court as a bootstrapped company, it settled (terms confidential) and took every API offline. Integrations built on it have been returning errors for over a year.

If you are still carrying dead Proxycurl code, two migration lessons matter. First, the outage pattern: a provider whose product depends on a single scraping method can disappear overnight, so favor multi-source architectures. Second, successors that replicate the scraping approach inherit the same legal exposure that ended Proxycurl. Our full Proxycurl alternatives guide walks through the shutdown timeline and a practical endpoint-by-endpoint migration path.

Crustdata (supplier credit)

Disclosure first: Crustdata is one of the providers inside the Cleanlist AI waterfall, so this entry is a supplier credit. It is unranked, and Cleanlist AI is not scored against Crustdata anywhere on this page.

Crustdata is a developer-first B2B data API. Per its published docs, it offers people and company screener endpoints alongside signal datasets: headcount growth over time (including by department), job openings, and funding data.

Hunter API (supplier credit)

Disclosure first: Hunter API is one of the providers inside the Cleanlist AI waterfall, so this entry is a supplier credit. It is unranked, and Cleanlist AI is not scored against Hunter anywhere on this page.

Hunter's API is built around email: Domain Search (who has emails at this company), Email Finder (name plus domain in, pattern-matched email out), and Email Verifier.

Pricing: Starter is $49 a month, or $34 a month billed yearly, with a free plan of 50 credits a month (hunter.io/pricing, read September 1, 2026).

Which People Search API Should You Choose?

Match the API to the actual job:

• You want cohort search plus verified, sendable contact data in one integration: Cleanlist AI API v2. Search free, enrich through the waterfall, sync to CRM, all behind one key. See pricing for credit math.
• You are licensing raw data to build your own product: People Data Labs. Budget for the separate Identify call and your own verification layer.
• Your team already runs on Apollo: use its API and accept the single-source accuracy ceiling.
• You want occasional individual lookups with phone numbers: RocketReach, if the top-tier API pricing fits.
• You are on Proxycurl: you are running on a dead API; migrate now.

And an honest note on where Cleanlist AI is not the answer: if you need the API on a $49/mo Starter plan (it requires Pro or above), if your architecture demands webhooks (v2 is poll-only), or if your cohort definitions depend on revenue, funding, or technographic filters (not currently in Cleanlist AI's people search), one of the other options, or a combination, will serve you better today.

Frequently Asked Questions

What is a people search API?

A people search API lets software query professional profiles with structured filters like job title, seniority, location, industry, and company size, and returns matching people as JSON. It differs from an enrichment API, which starts from a known identifier such as an email or LinkedIn URL and fills in missing fields. Modern providers increasingly combine both, so you can search for a cohort and then enrich the matches in the same integration.

What is the best people search API for developers in 2026?

For most developer teams, Cleanlist AI API v2 is the strongest option: 20+ search filters, free search reads, verified contact data through a 25+ provider waterfall (98% emails, 85% phones), and an MCP server so AI agents can drive it safely. People Data Labs is better if you specifically want to license one large raw dataset, and Apollo makes sense for teams already paying for its platform. The right answer depends on whether you want finished, verified contacts or raw data to build on.

Do people search APIs include email addresses and phone numbers?

Not always, and this is the biggest pricing trap in the category. People Data Labs bills email and phone through a separate, pricier Identify call on top of search. Apollo and RocketReach include contact data but meter it with credits or per-lookup pricing. Cleanlist AI includes verified email and phone in the enrichment step of the same call chain, at 1 credit for email, 10 for phone, and 11 for both.

Is the Proxycurl people search API still available?

No. Proxycurl shut down permanently on July 4, 2025 after LinkedIn sued its parent company Nubela over scraping profiles with fake accounts, and all of its APIs are offline. Any integration still pointing at Proxycurl endpoints is broken and needs to migrate. Replacements built on the same LinkedIn-scraping approach carry the same legal exposure, so most teams migrate to licensed, multi-provider APIs instead.

Does Cleanlist AI have a people search API?

Yes. POST /search/people on Cleanlist AI API v2 runs structured people search with 20+ filters, and searching is free; credits are only charged when you save results to a list or enrich them through the 25+ provider waterfall. The API uses clapi_ Bearer keys with granular OAuth scopes, returns async work as a workflow_id you poll, and requires a signed cost quote before paid bulk jobs. API access is included on Pro and Enterprise plans, and every endpoint is also exposed as a tool on the Cleanlist AI MCP server for AI agents.

Related reading

• Best B2B Data Enrichment APIs (2026)
• B2B Data Enrichment Accuracy Benchmark
• People Data Labs Alternatives
• Proxycurl Alternatives After the Shutdown