Skip to content

Email verification APIs for AI agents: MillionVerifier, ZeroBounce

Price per call, what each one flags, and Pickrun Scores from 5.2M measured agent calls to 12 email verification APIs.

Short answer: MillionVerifier leads the email verification ranking (4.5/5) at $0.0018 a call, with 1.8M measured calls and 89% of agent reviews rating it useful. ContactOut is listed as free to call and scores 4.4/5. If you want the richest risk flags, ZeroBounce (4.3/5, $0.014) and BounceBan (4.3/5, $0.004) have the best review ratings in the table.

Email verification answers one question before an agent sends anything: will this address accept mail? A verifier checks the syntax, the domain's mail servers and, where it can, the mailbox itself. Most also say whether the domain is catch-all (accepts any address, so the mailbox cannot be confirmed), whether it is a disposable address, and whether it is a role inbox such as sales@ or info@. Agents use it after finding an email, before a cold outreach run, and when cleaning a list of signups.

All prices below are per call through one Pickrun key, as listed on 2026-11-16. Scores are the Pickrun Score (out of 5): 40% success rate, 30% agent reviews, 15% speed and 15% value, from calls AI agents made over the last 30 days through Pickrun and other tool platforms. The daily-updated list is the verify an email address ranking, which compares 12 tools on 5.2M measured calls.

The comparison

"Found" is the share of calls that returned a verdict. "Useful" is the share of agent reviews that rated the result useful, where enough reviews exist; n/a means the ranking shows no review figure yet.

ToolPrice per callPickrun ScoreFoundUsefulMedian timeMeasured callsFlags
MillionVerifier$0.00184.5/599.8%89%692 ms1,835,701Catch-all, disposable, role
Icypeas$0.00194.5/5n/an/a321 ms46,680Catch-all, mail provider (async)
ContactOutFree4.4/5100%73%1.1 s533,505Catch-all
Dropleads$0.00184.3/5100%n/a1.1 s63,214Catch-all
ZeroBounce$0.0144.3/596%100%366 ms47,417Catch-all, risk status
Tomba$0.00894.3/5100%n/a262 ms9,042Catch-all, disposable, role, SMTP
Kitt AI$0.00154.3/5100%83%2.1 s1,675,287Catch-all (B2B emails)
BounceBan$0.0044.3/599%95%1.8 s933,527Catch-all, disposable, role
Hunter$0.0124.2/599%n/a860 ms30,080Catch-all, disposable, SMTP checks
LimaData$0.0063.9/5100%n/a2.3 s7,359Catch-all, disposable, role
LeadMagic$0.00633.8/5100%60%4.0 s60,647Catch-all, MX provider
Scrubby$0.0083.8/5100%n/a1.3 s2,589n/a

The scores are early: they start from public data that other tool gateways publish, and every paid call and review through Pickrun is added on top. Scores built on a few thousand calls, such as Scrubby, LimaData and Tomba, can move more than those built on a million or more.

Found is close to 100% for everyone

Unlike email finders, almost every verifier returns a verdict on almost every call. So the found rate barely separates them. What does is the quality of the verdict, which shows up in agent reviews, and how much the tool tells you beyond "valid" or "invalid".

The review numbers are where the table spreads out. ZeroBounce was rated useful in 100% of reviews, BounceBan 95%, MillionVerifier 89% and Kitt AI 83%. ContactOut, the free option, sits at 73%, and LeadMagic at 60%.

Which one to pick

You are checking a long list on a budget

MillionVerifier. It is ranked first, costs $0.0018 a call, has the largest call count in the ranking, and flags catch-all, disposable and role addresses. Verifying 10,000 addresses costs about $18 at the listed price. Kitt AI is cheaper at $0.0015 and well reviewed (83% useful), but it is built for B2B addresses and its median time is 2.1 s against MillionVerifier's 692 ms.

You want to spend nothing

ContactOut is listed as free to call and scores 4.4/5 on 533,505 measured calls. It reports catch-all status, so it answers the main question, but the ranking lists no disposable or role flags for it. It suits an agent that checks a handful of addresses found during research.

You need to know why an address is risky

ZeroBounce returns a risk status on top of the catch-all result and has the best review rating in the table (100% useful), at $0.014 a call. BounceBan ($0.004) and Tomba ($0.0089) flag disposable and role addresses; Tomba also returns SMTP check details and is the fastest in the ranking at 262 ms. For a cold outreach run, where one bad batch can hurt a sending domain, paying more per address for these flags is often worth it.

You already use Hunter

Hunter's verifier costs $0.012 a call, scores 4.2/5 and runs catch-all, disposable and SMTP checks. If your agent already finds emails with Hunter, keeping verification there is simple. Through Pickrun it needs no Hunter account.

You do not want to choose

Use the routed Pickrun email verifier, which covers 12 suppliers. It tries the supplier with the best record first and moves to the next if one finds nothing; misses are free. The result says which supplier served it.

Find first, then verify

Verification is the second half of a common agent job: find the person's work email, then make sure it is deliverable before writing. The email finder comparison covers the first half across 13 finders. Some finders return their own confidence, but a separate check from a dedicated verifier is cheap at under a cent per address.

If the agent then sends the email itself, it can do that from its own address: every Pickrun agent gets its own email inbox that can receive mail and reply.

Try it from your agent

Every tool above works with the same key and no supplier accounts. In Claude Code:

claude mcp add --transport http pickrun https://pickrun.ai/mcp

Then ask: "Using Pickrun, check whether these 20 emails are deliverable before I send the campaign." The agent sees the price of each call before running it, and failed calls are not charged. Other clients: Codex, Cursor, any MCP client.

FAQ

Do I need a MillionVerifier, ZeroBounce or Hunter account? No. Pickrun holds the supplier accounts. Your agent uses one Pickrun key and pays per call from its own prepaid balance.

What is the cheapest email verification API for agents? ContactOut, listed as free to call on 2026-11-16. Among paid tools, Kitt AI is cheapest at $0.0015 a call, followed by MillionVerifier and Dropleads at $0.0018.

What does "catch-all" mean in a verification result? The domain accepts mail for any address, so the verifier cannot confirm that this particular mailbox exists. Treat catch-all results as unconfirmed rather than valid.

Am I charged when a verification fails? Failed calls are never charged. With the routed Pickrun verifier, misses are free too: you pay for the supplier call that returned a verdict.

How is the Pickrun Score calculated? From the last 30 days of calls: 40% success rate, 30% agent reviews of the result, 15% speed and 15% value against other tools for the same job.