Company Lookup API Pricing: Models, Rates and a Cost Worksheet

Company lookup API pricing comes in four shapes: pay-per-call, credits, seats and enterprise contracts. The number that matters is cost per correct, found record, which is always higher than the headline rate because some lookups return nothing and some return something wrong. Work that number out from a pilot, not from a price page.

The four pricing models

Model	How it charges	Fits	Watch for
Pay-per-call, prepaid	A fixed amount per call or per found result, from a balance	Uneven or modest volume, tests, developers	Whether misses are billed
Credits	Credits per call or per field, bought in packs	Mixed data needs	Expiry, different credit costs per data type
Per seat	Monthly fee per user, with credit allowances	Sales teams working in an app	Seat minimums, overages, annual contracts
Enterprise licence	Quoted annually	Large organisations	Minimums, auto-renewal, bundles you do not use

None is better in general. Pay-per-call suits a team that wants to try a day of data. Seats suit a sales team that lives inside a tool. Contracts suit buyers who need support and procurement paperwork.

A concrete example: BusinessMCP

BusinessMCP’s company data API publishes its rates:

• Company lookup: $0.02 per call, charged only when a company is found.
• Bulk lookup: $0.02 per item, up to 100 domains per call, charged for companies found.
• Website status: $0.005 per call, when found.
• Email pattern: $0.01 per call, when found.
• Company search: $0.01 per company returned.

It is prepaid with no subscription, and new accounts get $5 free for 7 days. Its comparison page adds that prepaid packs start from $10. Per its own reference pages, a charge refused for low balance is never billed later, and calls that fail on the vendor side are refunded.

At $0.02 each, the $5 trial covers up to 250 found companies, which is enough for a decent pilot. That is arithmetic from the published price, not a separate claim from the vendor.

How other vendors charge

The same comparison page, which is vendor-published and dated September to October 2026, lists the following. Verify each on the vendor’s own site.

Vendor	Stated model
ZoomInfo	About $0.38 per credit at the $1,000 tier; credits expire after 12 months
Apollo.io	$49 to $119 per user per month on annual billing, with credit use per enrichment
Hunter.io	Free 50 credits a month; $49 to $299 a month for 2,000 to 25,000 credits
Crunchbase	API sold as an enterprise licence
Coresignal	$49 to $5,000 a month by credit tier
Clearbit	Folded into HubSpot plans; no new standalone API keys reported

These tools are not like for like. ZoomInfo and Apollo bundle people and intent data; Hunter specialises in email finding; Crunchbase focuses on funding. A domain-to-firmographics call is a narrower product, so a lower price per call is partly a narrower scope. The page itself notes that BusinessMCP’s per-record price is not the lowest at volume.

Key takeaway: compare cost per useful answer for the data you actually need, not the lowest headline number across different products.

A cost worksheet

1. Count unique domains you will look up per month after cleaning and deduplicating. Call it D.
2. Estimate the hit rate h from a pilot of a few hundred domains. Do not use a vendor’s claim.
3. Compute found records: D × h.
4. For pay-per-found pricing: cost = D × h × price.
5. For per-call-regardless pricing: cost = D × price.
6. For seats: cost = seats × monthly fee, then check whether the credit allowance covers D.
7. Add a 20 percent buffer for retries, re-enrichment and growth.

Example with illustrative numbers: D = 5,000, h = 55 percent, price $0.02 found. Found = 2,750; cost = $55 before buffer. The 55 percent is invented for the example. Your hit rate may differ widely, which is why step two matters most.

Hidden costs

• Misses billed. Always test with a fake domain and check your balance before and after.
• Credit expiry. Unused credits that lapse after 12 months are a real cost if your usage is lumpy.
• Re-enrichment. Data goes stale, so budget for refreshing important accounts.
• Engineering time. A cheap API that needs a week of integration may cost more than a pricier one with an SDK.
• Overlap. If you already pay for HubSpot, its enrichment may cover your needs. BusinessMCP’s Clearbit page explicitly advises HubSpot users not to pay twice.

Controls worth having

Ask whether you can cap spend per API key, set a low-balance alert and see per-call usage. BusinessMCP’s error reference mentions a key cap and insufficient balance as reasons a charge is refused, which implies per-key limits exist. A spending cap is especially useful when a script runs unattended; see the bulk lookup guide.

Choosing a model

• Under a few hundred lookups a month: use a free tool such as the free company lookup and revisit when time costs more than data. See free company lookup tools.
• Thousands a month, steady: price pay-per-call against a tier and pick the cheaper.
• Sales team needing contacts: compare seat platforms, and treat company data as a bonus.
• Developers building a product feature: prefer prepaid, documented rates and clear errors; the curl walkthrough shows what that looks like.

Prices in this market move fast. Treat any table, including this one, as a starting point and confirm on the vendor’s page, such as BusinessMCP’s pricing page, before you commit budget.

Worked scenarios

Scenario A: a quarterly CRM clean. You have 12,000 contacts but only 4,000 distinct company domains once duplicates and free-mail addresses are removed. You run them once per quarter. Under a pay-per-found model the spend scales with finds only, and if you refresh the same accounts every quarter you should still budget for the full list each time, because stale records are the reason you are re-running it. The cost is lumpy, with nothing in between runs, which suits prepaid balances well and subscriptions poorly.

Scenario B: a signup form. You enrich every new work-email signup in real time. Volume is steady and small, perhaps a few hundred a month. A single lookup per signup at a per-found price is easy to forecast, and the important engineering work is handling a miss gracefully: show the form as normal and move on. Do not block a signup because the lookup returned nothing.

Scenario C: an outbound team of ten. Here the data is a small part of a bigger toolkit. A seat platform with contacts, sequences and intent may be priced per user, and the real question is whether you use enough of it. If the team only needs company facts, a per-call API alongside a cheaper sequencing tool may cost less than paying for ten seats of features nobody opens.

Questions to put to any vendor about price

1. Is a call that finds nothing billed, refunded or free?
2. Are there minimum top-ups, expiry dates or auto-renewals?
3. Can I cap spend per key and see usage per call?
4. What are the rate limits, and do heavier endpoints have lower ones?
5. Are failed calls on your side refunded automatically?
6. What changes on price after a trial ends?

Get the answers in writing, or from published documentation, rather than from a sales call alone. A vendor with clear published error and billing behaviour makes your forecast far more reliable.