Skip to content

Changelog

What is new in the PricklyMails email verification API. Follow along as we ship.

October 2026

  • Check each Google Forms response. A short script writes the recommendation for each new response's address beside it in the response sheet.
  • Verify from n8n or Make. A page for each platform: keep your key as a credential, verify the address a form brings in one request, or a whole list as a batch.
  • Verify from Claude, ChatGPT or Cursor. A page per assistant shows how to connect it to the PricklyMails MCP server, with one click where it allows it, then which agent to approve and what to ask.
  • Refuse a dead address at signup. New pages for Supabase, Firebase, Auth0 and WordPress show how each one's own signup hook can ask PricklyMails about a new address and refuse the account when the address cannot receive mail, with the code to paste and a free test key to try it.
  • Get a result in the call that asks for it. Send Prefer: wait=4 with POST /api/verify, and the response waits up to that many seconds, 12 at most: a verification that finishes in time comes back with its result, one that takes longer is answered as before, so a signup form can check an address in one request.
  • A free email bounce calculator. Enter your list size, your sending platform and a share of dead addresses: the calculator shows what they cost you, a higher contact tier or a campaign a suspension would stop, its revenue estimated from a published benchmark until you enter yours, beside what verifying the list costs, and which published thresholds the next send crosses at Brevo, Klaviyo, HubSpot, Mailjet or Amazon SES. It sits under the new "Free tools" menu.
  • Twenty-six new step-by-step pages. Kit, lemlist, Loops, Substack, Campaign Monitor, Ghost, Shopify, PrestaShop, Notion, Google Sheets, Airtable, Odoo, Resend, Gmail, Outlook, Intercom, Wix, Squarespace, BigCommerce, Mailtrap, Gumroad, Lemon Squeezy, Stripe, Paddle, Clerk and ClickUp: export your list, verify it, and act on each result in the tool itself.
  • Automate it with n8n or Make. A new docs page walks through the PricklyMails side of a workflow, request by request.
  • The cause behind each result. A verification's detail now says exactly why, "The domain publishes a null MX" under "Domain refuses all email", with the reason code the API returns.
  • A Mailjet page. Export your list from Mailjet, verify it, then add the addresses to remove to your campaign exclusion list, each step sourced from Mailjet's help.
  • Plain words in the dashboard. Checks and reasons say what they mean: "Mailbox exists" rather than "SMTP probe", "Domain receives email" rather than "MX lookup", "Generic address" rather than "role-based". A batch's detail starts with what to send, review and remove.
  • The confirmation step names its sender. After signing up, the code step says the account is created and which address the code comes from, so it is easy to find in a spam folder.
  • Verify an address from "New verification". The form opens in a window, you choose the checks that run, and a check outside your plan names the plan that unlocks it. The history shows each verification's address, in the browser that ran it.
  • The API reference is complete again. The OpenAPI specification had lost most of its route descriptions and schema details; they are back, and a check now keeps them from going missing.
  • Choose the checks of a single verification through the API.POST /api/verify takes an optional checks list, as a batch already does: those checks run instead of the key's own, within your plan.
  • Batch options fold away. A batch's name and checks move under "Options", and a batch with no name takes its file's.
  • A batch holds at least 20 addresses. Below that, verify addresses one at a time: the dashboard, the API and the MCP now refuse a batch of fewer than 20 distinct valid addresses. A batch sent in chunks may start smaller, and only closes once it reaches the minimum.
  • Already verified Gmail, Outlook and Yahoo addresses come from the cache. Verifying again an address at these providers, or on a domain that cannot receive mail, reuses the recent result instead of starting over, and costs nothing, like any address served from the cache.
  • HubSpot and Klaviyo pages, and a guide to what dead addresses cost. Export, verify, then bring the result back into HubSpot or Klaviyo, each step sourced from their official help; and what dead addresses really cost, from the rules email tools publish.
  • The mailbox check shows its real result. In a verification's detail, an accepted mailbox showed as a failed check; it now shows as passed, and a server that did not answer shows as "No answer".
  • Batches counted the way they are billed. Duplicates and invalid lines are set apart before you send, the estimate counts only what you pay for, and a list your balance cannot cover is stopped with the number of tokens missing.
  • Addresses in batch results. The results table shows each address next to its verdict, in the browser that sent the batch.
  • A Notifications page. What happens on your account, in one place: API keys, webhooks, agents, finished batches, billing and security.
  • Messages that are easier to read. Identical messages are grouped, the stack pauses while you hover it, and Escape clears it.
  • Picking a plan takes you to checkout, even without an account. The plan you choose on the pricing page stays with you through sign-up, email confirmation and sign-in, up to the payment page.
  • 14 days to change your mind. At checkout, you ask to use your tokens right away. For 14 days after the order, you can withdraw: the tokens you have not used are refunded, in proportion to the price you paid.
  • An email after every payment. What you bought, the amount, the date, the tokens credited, and how to withdraw.
  • Emails in your language. Every email goes out in English or French, in the language you use the site in, and its links open the pages in that language.
  • Verify an address from the dashboard. No API key needed to try an address: type it, and the verdict shows up.
  • Free batches of up to 100 addresses. The Free plan goes from 10 to 100 addresses per batch, and the batches page shows your plan's limit.
  • Catch-all detection, only when you ask for it. It is part of the Pro plan and above, and runs only when you request it. Without it, a confirmed mailbox scores 88; with it, 96 when the domain does not accept everything.
  • Your webhooks get every event. Every active endpoint on your account receives the events, not only the newest one.
  • Addresses with an apostrophe are accepted. An address such as o'brien@example.com is no longer reported invalid.
  • A clearer batch report. Malformed addresses are listed with the Remove recommendation, the values are in your dashboard's language, and accents display correctly in Excel.
  • Guides and tool pages. Guides on bounces, list cleaning and choosing a verifier, and step-by-step pages for Brevo and Mailchimp, each step sourced from the tool's official help.
  • Top-ups from €4. The smallest top-up goes from 1,000 to 4,000 tokens, or €4.
  • Sign-in stays in your language. A link from an email, two-factor verification, and the return after an expired session keep you on the pages in your language, and an expired verification session now shows a message.
  • The dashboard opens from a link in an email. Already signed in, a link to the sign-in page takes you straight to your dashboard: it could open empty, and the sign-in form showed for an instant first.

September 2026

  • The MCP server on the Free plan. One AI agent can verify one address at a time and read your results, on your free tokens.
  • Two-factor codes that are easier to enter. Six boxes, a code pasted in one go, and your phone's autofill.
  • Forgot your password without retyping your email. The address you typed at sign-in carries over to the reset form.
  • Sign-up checks your address with the real engine. The address you sign up with goes through the same verification as the API. One that cannot receive email, or a disposable one, is refused before the account exists; an uncertain result never blocks you.
  • Your addresses never touch the disk. An address waiting to be verified stays in memory while it is verified, never written to disk. Our DNS lookups go through our own resolver, with no third-party service.
  • A deleted account is erased for good. 28 days after deletion, the account's history (verifications, keys, webhooks, sessions) is permanently erased. Only the payment records the law requires us to keep remain, without your email.
  • Access cut off at once. Banning or deleting an account immediately revokes all its API keys and agents.
  • An honest verdict when a provider refuses verification. Some providers, such as GMX, web.de or IONOS, refuse verification before looking at the mailbox. The address comes back "not verifiable", with the reason, and that check is refunded.
  • A data processing agreement. Our DPA, under GDPR article 28, is published at /dpa.
  • A new landing page. What we check, the result of a real list, our real figures, and the questions people ask us, in the order you ask them.
  • Fairer charts. The activity and balance charts no longer show one period too many, and offer 8, 28 and 88-day windows.
  • The password reset link goes to the right place. It used to open a page that did not exist.
  • Connect an AI agent over OAuth. Create scoped agents and authorize an MCP client, like Claude, to act on your account with only the permissions you grant, over OAuth 2.1 with PKCE. The agent verifies addresses with your own API keys, and you can revoke it at any time.
  • Hosted MCP server. PricklyMails is now a remote Model Context Protocol server at mcp.pricklymails.com. Add it as a connector and your agent gets verification, balance, and usage as tools, no local setup.
  • A full agent toolbox. The MCP server now also covers bulk verification, API-key list and revoke, webhook setup, wallet history, and job listing, so an agent can run a whole verification workflow over its scoped connection, not just one address at a time.
  • SMTP on every plan. The live SMTP mailbox check is now included on every tier, including Free. Catch-all detection becomes the premium check on Pro and above.
  • Built for developers and AI agents. The landing page now positions PricklyMails as a high-speed email verification API for both developers and AI agents, with a new section on using it as an MCP tool so an agent can verify an address before sending.
  • Live numbers on the homepage. The landing stats now show real figures, total verifications and average latency, served from a public cached endpoint, replacing the placeholder tiles.
  • Signup email check. When you register, we run a quick, non-blocking check on your address and show a heads-up if it looks invalid or disposable. It never blocks you from signing up.
  • Cached scores follow the latest model. When we recalibrate scoring, a previously cached result is re-graded under the current model on read, so an older result never shows a stale score.
  • Webhooks on Starter and above. Webhook endpoints are now part of the Starter plan and up. The Free plan shows the webhooks tab with an upgrade prompt.

August 2026

  • Fewer false positives on unreachable servers. When an SMTP check times out, cannot connect, or returns an ambiguous reply, the address is now reported as unknown (we could not confirm it) rather than being optimistically marked valid, and an inconclusive catch-all probe no longer inflates the confidence score.
  • Bulk verification. Drop a CSV or paste a list and verify up to a million addresses in one batch. Everything is parsed in your browser before anything is sent, and you get per-address results as they come in.
  • Sharper deliverability signals. Greylisting (a mail server asking us to try again later) is now detected and retried instead of guessed, and big accept-all providers like Gmail, Outlook, and Yahoo are reported as risky rather than falsely marked valid.
  • Recalibrated quality score. The 0-100 quality score is now graded within each deliverability status, so it never contradicts the verdict.
  • Sandbox test mode.pm_test_ keys now run a deterministic sandbox: no SMTP, no billing, and canned outcomes chosen from the address local part (valid@, invalid@, catchall@, disposable@, role@), so you can build and test an integration without spending credits.
  • Account recovery. Request a fresh confirmation link when the original one expired, and recover from an expired link instead of getting stuck on a dead-end screen.
  • OpenAPI specification. Every endpoint is now described by a generated OpenAPI 3.1 document, served at /api/openapi.json and /api/openapi.yaml, so you can import the whole API into Postman or Insomnia in one step.

July 2026

  • Webhooks. Register an endpoint and receive a signed notification the moment a verification job completes, instead of polling for the result.
  • Single-email verification. Submit an address to /api/verify and get syntax, MX, SMTP, disposable, role, free-provider, catch-all, and typo checks with a deliverability verdict and a quality score.
  • Dashboard. Manage everything from one place: an overview with usage charts, your verification history with a per-check breakdown, API key management, active sessions, and your token wallet.
  • Two-factor authentication and account management. Turn on TOTP two-factor from your profile, review and revoke your active sessions, and delete your own account, each confirmed with your current password.
  • Try it live. A public demo on the landing page runs a real verification (syntax, MX, disposable, and typo checks) with no signup, so you can see a result before creating an account.

June 2026

  • Accounts and authentication. Create an account, confirm your email, sign in, reset a forgotten password, change your email or password, and protect your login with two-factor codes. Every confirmation and notification arrives as a branded transactional email.
  • API keys. Generate live and test keys (pm_live_ and pm_test_), scope each key to the exact checks it may run, and manage them from your dashboard. Each key carries its own plan-driven rate limit.
  • Wallet and token billing. Usage is metered against a token wallet: every verification reserves its cost up front, then commits or refunds it based on which checks actually ran. Checks that never ran are refunded automatically.
  • Subscription plans and Stripe checkout. Pick a plan, check out through Stripe, and manage your subscription from the Stripe billing portal. Buy extra tokens any time with a dynamic per-token top-up.
  • Privacy by design. Verified email addresses are never stored: only salted hashes are kept, so your lists never sit in our database in the clear.

April 2026

  • Early access. The PricklyMails landing page and waitlist went live, so you could reserve your spot while the verification API was still in development.