Website Data API Pricing: What Per-Domain Checks Cost

Website data APIs are usually priced per call, prepaid. On BusinessMCP the AI readiness grade costs $0.03 per domain and each of the other three checks costs $0.01, so a full four-check audit of one domain is $0.06. Auditing 100 domains that way costs about $6, before any retries or repeat runs.

The published rates

These figures come from BusinessMCP’s own pages for its website checks API and the data API overview. They can change, so confirm before you budget.

Check	Price per call
AI readiness grade	$0.03
AI crawler access	$0.01
WebMCP check	$0.01
Email authentication	$0.01
All four on one domain	$0.06

The service is prepaid with no subscription. That matters for planning: there is no seat fee or minimum monthly spend described, so cost scales directly with how many calls you make.

Budget scenarios

Multiply domains by checks by runs per year. All figures below use the rates above and assume every call succeeds.

Scenario	Calls per run	Cost per run	Runs per year	Yearly cost
One site, four checks, weekly	4	$0.06	52	$3.12
25 client domains, readiness only, monthly	25	$0.75	12	$9.00
100 domains, four checks, monthly	400	$6.00	12	$72.00
500 domains, email auth only, weekly	500	$5.00	52	$260.00

The takeaway is that audit-style usage is cheap in absolute terms. The number that usually surprises people is not the price per call but how quickly a naive schedule multiplies it. Checking 500 domains daily instead of weekly is 7 times the spend for data that rarely changes.

Key takeaway: match the check frequency to how often the underlying thing changes. Email DNS records and robots.txt change rarely, so weekly or monthly is enough for most portfolios.

What the free credit covers

The free trial is $5, once, valid for 7 days, according to BusinessMCP’s data API page. At the published rates that is roughly 166 AI readiness calls or 500 of the $0.01 checks, or any mix. The credit applies to BusinessMCP’s own tools and, per the vendor, does not cover people data. The trial starts when you save a card on the Billing page; the card is not charged, and the vendor says there is one trial per person and per card.

Use it deliberately. A good test plan for $5:

1. Run all four checks on five domains you control (30 cents).
2. Include one domain that is deliberately misconfigured, such as one with no DMARC record.
3. Include one domain with a blocked or missing robots.txt.
4. Log every raw response before parsing.
5. Decide from the results whether the checks answer your questions.

If you want to see the behaviour without writing code first, the same checks are available as free web tools, for example the AI crawler checker.

Costs that are not on the price list

Price per call is the visible number. Three others deserve a line in your estimate.

Engineering time. A script that loops over domains, stores results and alerts on change is small, but it is not zero. See the Python example for the size of the job.

Retries and unknowns. A lookup that comes back unknown may deserve a retry later. Whether an unknown result is billed depends on the vendor’s terms, so read them rather than assume. BusinessMCP says unreadable sites are refunded rather than graded.

Alert fatigue. The expensive failure is not the $0.01; it is a noisy alert channel that people mute. Design thresholds before you scale up, as covered in scheduled website monitoring.

Comparing with other pricing models

Other vendors in adjacent categories sell monthly plans with call allowances, enterprise contracts or credits that expire. Per-call prepaid pricing suits irregular, bursty audits, such as an agency onboarding a client or a quarterly review. A flat monthly plan can win for steady, heavy volume. To compare fairly, convert every offer to cost per successful check at your real volume, and include any minimum commitments.

For a larger run the maths and the workflow come together in bulk domain audits.

Safe-guard against surprise spend

• Keep a hard cap in your own code: stop after N calls per run.
• Deduplicate and normalise domains before sending them, since Example.com and example.com are the same site.
• Cache results for as long as the data is plausibly unchanged.
• Alert yourself on any HTTP 402, because it means the balance is empty and your monitoring has quietly stopped.

Pricing, free-credit terms and refund rules are the vendor’s to set, so verify them on the official pages before you rely on any number here.

How to turn a price list into a decision

A per-call rate is only meaningful against a number of calls you actually expect, so work through a short worksheet before committing.

1. Count domains. Separate live sites, sending-only domains and parked domains. They need different checks.
2. Assign checks. For each class, list the checks you will run. A sending-only domain may need only the $0.01 email check.
3. Pick a cadence. Monthly for stable sites, weekly for sending domains, daily only during migrations.
4. Multiply. Domains times checks times runs per year times price per call.
5. Add a margin. Reserve 20 to 30 percent for retries, new domains and ad hoc checks.
6. Compare against the alternative. Estimate the engineering hours of a DIY script and the free tools’ limits for the same coverage.

Here is a fully worked example. An agency manages 40 client sites and 15 sending domains. The sites get readiness plus crawler access monthly: 40 times $0.04 times 12 equals $19.20. The sending domains get the email check weekly: 15 times $0.01 times 52 equals $7.80. Total about $27 a year before margin. Even doubling it for retries and ad hoc checks leaves a figure that is trivial next to the cost of one missed DMARC problem.

When the per-call model is a poor fit

Per-call prepaid pricing is best for bursty or modest volume. It is a weaker fit if you need very high daily volume, because there may be cheaper committed-volume plans elsewhere, or if finance requires a fixed monthly invoice for budgeting. In either case ask the vendor whether committed pricing exists; the pages we reviewed do not describe one, so do not assume it does.

It is also a poor fit if you want the cheapest possible check of a handful of domains once. For that, the free browser tools are enough, and BusinessMCP’s own AI readiness grader is one of them, subject to its per-visitor rate limiting.

Reading the fine print

Check these items in the vendor’s terms before relying on a budget:

• Whether unused credit expires. The trial credit does, after 7 days; paid balance terms should be confirmed separately.
• Whether refunds apply to unreadable sites, and how quickly.
• Whether the same price applies via REST and via MCP. BusinessMCP states the same name, input and price for both.
• Whether taxes are added at checkout.
• Whether any endpoint is billed only when a result is found. On the company and IP APIs BusinessMCP charges only when found, whereas the website checks are listed per call, so do not carry assumptions across products.

The last point is a good example of why you should price each product separately: a company lookup that bills only on a hit behaves very differently from a check billed on every call.

Quick reference

Question	Answer from the vendor pages
Subscription needed?	No, prepaid
Free credit	$5, once, 7 days
Empty balance behaviour	HTTP 402, no later charge
Cheapest check	$0.01
Most expensive website check	$0.03

If you want to see what a call looks like before sizing a budget, the Python example is the shortest path from key to first response.