Integration
CampaignKit runs a hosted email verification MCP server that Claude connects to with your API key, so there is nothing to install. Ask Claude in plain language to check an address or a list, and you get back a score from 0 to 10 and a classifier: valid, risky or invalid.
Connect Claude to CampaignKit and you can verify emails from a chat: paste an address or a list of leads, and Claude checks them with our email verification MCP server and tells you which ones are safe to send to.
CampaignKit's email validation MCP server is a remote MCP server at https://api.campaignkit.cc/v1/mcp/sse. We built and run it ourselves; it doesn't go through Zapier, n8n or another vendor's API. Claude reaches it by URL with your CampaignKit API key, and so can other clients that support remote MCP servers. It is not made, listed or endorsed by Anthropic: you add it to Claude yourself, as you would any custom MCP server.
You can verify emails with Claude in plain language. Some prompts that work well, and the tool Claude uses for each:
validate_email.create_validation_job with the list.get_job_status, then get_job_results.get_validation_stats.validate_email.| Tool | What it does | Credits |
|---|---|---|
validate_email |
Validates one email address and returns its score and classifier. | 1 per validation |
create_validation_job |
Starts a batch job for a list of addresses, with an optional label. | 1 per email, charged during processing |
get_job_status |
Shows a job's state (pending, running, done, failed or paused) and a summary of deliverable, undeliverable and risky addresses. | Free |
get_job_results |
Returns a job's results, 50 per page. Claude can ask for the next page. | Free |
get_validation_stats |
Shows your account's validation stats for the last 7 days by default, up to 365 days. | Free |
The MCP server reference lists every parameter and response field.
Setup takes a few minutes:
ck_. Keep it private, like a password.https://api.campaignkit.cc/v1/mcp/sse to your MCP client, and send your key as a Bearer token in the Authorization header (Bearer ck_…).For the exact configuration and troubleshooting, follow the Claude AI integration guide.
Every address comes back with a score from 0 to 10 and a classifier. Claude puts them into a plain answer. For example:
You: Can you check if john.doe@company.com is valid?
Claude: The email john.doe@company.com has a score of 10/10 and is classified as "valid".
How to read the score:
| Score | Meaning |
|---|---|
| 10 | Fully verified: the mailbox exists. |
| 9 | Valid, but the mailbox couldn't be confirmed. |
| 2 | Usable but risky. |
| 0 | Will bounce. |
The classifier is valid (deliverable), risky or invalid (will bounce). A mailbox can't always be confirmed, so some real addresses score 9 rather than 10. Risky addresses need your judgment: review them before you send, or send to them with care.
Typo suggestions. When an address fails and its domain looks like a well-known provider, the result includes a suggestion. For john@gmai.com, which scores 0 and is invalid, the suggestion is john@gmail.com. When no suggestion applies, there is none. Show the suggestion to the person who entered the address instead of rewriting it silently.
For a batch job, Claude can summarize the job and then list the results page by page. Treat the numbers in any example conversation as examples; your results depend on your list.
Other email verification MCP servers exist. Some run as a local process on your machine; others are hosted connectors from third-party platforms. As of October 2026, from each project's listing or README:
| CampaignKit | MCP Email Verify | BounceProtect | |
|---|---|---|---|
| Where it runs | Hosted by CampaignKit, reached by URL | Locally, set up on your machine | Locally, as a STDIO server |
| What you supply | A CampaignKit API key | Python 3.11+, UV and your own AbstractAPI key | Your own BounceProtect API key |
| Bulk checks | Batch jobs with status and paged results | None listed in its README | validate_emails_bulk, up to 100 addresses |
| Free credits | 200 validations | Not stated | 100 credits |
A local server can suit you if you want to run everything on your own machine. Platforms such as Composio also host connectors for other verification services: you add their connector URL to Claude, then authorize the service. Check each project's page for current details.
Every new account gets 200 free validations, no credit card required. After that, you pay as you go from $6 for 2,500 emails ($2.40 per 1,000), with no subscription, and credits never expire.
Through Claude, each validation costs 1 credit, and batch jobs charge 1 credit per email during processing. Checking a job's status, reading results and viewing stats are free. Credits are only deducted for addresses with valid syntax. If your credits run out, Claude gets an "insufficient credits" error, and you can top up in the app. See pricing for all packs.
Claude is a trademark of Anthropic. CampaignKit is not affiliated with or endorsed by Anthropic.
Start with 200 free validations. Upgrade only when you're ready.
No credit card required • Cancel anytime