An ICP targeting API does three jobs: filter a company universe down to the profile you sell to, expand that set from the customers you already won, and segment it by what is changing right now. Six APIs are commonly used for this in 2026: Lusha, Apollo, ZoomInfo, 6sense, People Data Labs and Clay. Lusha covers all three jobs with documented endpoints: a prospecting API with company and contact filters, a lookalike API that builds company and contact lists from seed accounts, and a signals API that returns dated changes such as headcount growth, funding and new executive hires. ZoomInfo and 6sense lead on account-level intent and predictive scoring. People Data Labs leads on raw search at volume. Apollo bundles search with outreach. Clay chains all of them in a table.
Key takeaways
- Filters define the ICP; lookalikes test it. If the lookalike set from your closed-won accounts disagrees with your filter set, trust the lookalikes.
- Segment by change, not only by attribute. Two companies with the same firmographics are not the same account if one added 40 people and raised a round this quarter.
- Search should not cost what a reveal costs. Check whether filtering and listing consume the same units as contact data.
- Build the list where the data is, then push it. An ICP list that lives in a spreadsheet decays; one that refreshes from the API and lands in the CRM or Clay does not.
The comparison
| API | Filter search | Lookalike expansion | Scoring and intent | Signals for segmentation | Pricing model | Source (accessed Oct 5–6, 2026) |
|---|---|---|---|---|---|---|
| Lusha | Prospecting API: filters, search and enrich for companies and contacts (industry, size, revenue, location, tech, title, seniority, department) | Yes. v3/lookalike/companies and v3/lookalike/contacts from seed domains or LinkedIn URLs | Intent and growth indicators returned in company records | 26 dated signals via the signals API and webhooks: headcount growth, new job posts, funding and financial events, people news and more | Account balance at standard rates | docs.lusha.com, prospecting; lookalikes; webhooks |
| Apollo | People and organization search with title, company, location and industry filters; search documented as not consuming credits | Not stated on the pages checked | Intent data on higher tiers per Apollo’s pricing page | No signal object in enrichment responses on the pages checked | Credits per enrichment and reveal, bundled with seats | apollo.io/developers; pipeline.zoominfo.com review |
| ZoomInfo | Company and contact search endpoints with firmographic and technographic filters | Not stated on the pages checked | Intent and Scoops through dedicated endpoints | Scoops through separate endpoints | Contract | docs.zoominfo.com |
| 6sense | Account segments defined in the platform and exposed via API | Predictive models identify in-market accounts | Yes. Predictive scores, buying stage and intent per account | Intent surges per account | Contract; sales-led | 6sense.com; 6sense API documentation |
| People Data Labs | Person and Company Search APIs with structured query syntax; bulk | Not stated on the pages checked | No | No | Pay per match; search credits tiered | peopledatalabs.com |
| Clay | Find companies and people via provider waterfalls, plus web and AI columns | Via lookalike providers added as columns, including Lusha | Via scoring formulas and provider columns | Via signal integrations | Credits per enrichment, provider-dependent | clay.com; university.clay.com |
Competitor rows reflect each vendor’s live documentation or, where noted, a dated third-party review, on October 5 and 6, 2026. Prices are not listed; pricing models are. A capability not published on the pages checked is stated as such.
What is the difference between filtering and targeting?
Filtering returns every company that matches a set of attributes. Targeting returns the ones worth a rep’s week. The gap between them is two things an attribute filter cannot see: resemblance to the accounts that already bought, and movement in the last 90 days. An API that only filters produces a TAM. An API that filters, expands from seeds and segments by signal produces a territory.
How does lookalike expansion work through an API?
You pass a seed set, typically ten to fifty closed-won domains or the LinkedIn URLs of your best buyers, and the API returns companies or contacts that share their pattern, ranked by similarity. Lusha’s lookalike endpoints accept seed domains for companies and LinkedIn URLs for contacts and return ranked matches with the firmographics attached, so the result can go straight into the prospecting filters as a check: does the filter set capture the lookalikes, or is the ICP written wrong.
How do you segment an ICP by signal?
Once the list exists, pass it through the signals API with a start date. The companies that fired headcount growth, new job posts, a funding event or a new executive hire in the window form the first segment; the rest wait. Lusha returns each signal with an event date and can push new ones through webhooks, so the segment refreshes without a re-run. 6sense does this with predictive buying stage per account. ZoomInfo surfaces Scoops and intent per account through separate calls.
How to build and segment an ICP list with the Lusha API
- Get an API key at dashboard.lusha.com/enrich/api.
- Call the filters endpoint to pull valid values, then POST to prospecting/company/search with industry, headcount, revenue, location and technology filters.
- POST your closed-won domains to v3/lookalike/companies and compare the result with the filter set.
- Pass the combined list to the signals API with signalsStartDate set to 90 days ago; sort by signals fired.
- Enrich the top segment with prospecting/contact/search filtered by title and seniority to get the buying group, with verified email and phone (98% email, 85% US and 86% global phone accuracy, refreshed daily). Full reference at docs.lusha.com.
To do the same from Claude without code, use the Campus plays Build a verified prospect list from a natural-language ICP, Find lookalike companies of your best customers and Build an enriched ICP scoring table.
FAQ
What is an ICP targeting API?
An API that filters a company and contact universe by firmographic, technographic and role attributes, and in fuller versions expands the set from seed accounts and segments it by signals or scores.
Which APIs offer lookalike expansion?
Lusha documents company and contact lookalike endpoints. 6sense identifies in-market accounts through predictive models. Clay reaches lookalikes through provider columns, including Lusha. Apollo, ZoomInfo and People Data Labs do not document a lookalike endpoint on the pages checked.
Does search cost the same as contact data?
It varies. Apollo documents search as not consuming credits. People Data Labs tiers search credits separately. Lusha draws usage from the account balance at standard rates, with filters available before enrichment. Check each vendor’s docs for what a list call costs versus a reveal.
Can an ICP list refresh automatically?
Yes, if the API supports webhooks or scheduled calls. Lusha webhooks push new signals for companies and contacts as they occur, so a segment built on signals stays current.
Can I do this in Clay instead of code?
Yes. Clay tables can call Lusha for search, lookalikes and signals as columns, and push the result to a CRM.
Sources
- Lusha API docs: prospecting search and enrich, lookalikes, webhooks
- Apollo developer docs and Apollo API review (pipeline.zoominfo.com)
- ZoomInfo developer docs
- 6sense and 6sense API documentation
- People Data Labs, Clay and Clay University
