- Email validation, hygiene-grade.
POST /api/validatechecks syntax, MX records, disposable domains and role accounts before you import or send, at 2 units per address. Deliberately not SMTP-grade verification, and the docs say exactly where it stops. - Fuzzy matching moved from the browser to the API. The in-browser checkbox is gone: the comparison is pairwise and froze the tab on real-world lists, which is worse than not offering it. It now runs server-side via
options.fuzzyin the API, with a similarity threshold, any API key including the free one, and candidate blocking in place of all-pairs comparison, so a request can carry up to 50,000 rows, the same ceiling as the paid per-request cap. The browser tools stay exact-only and instant at any size. - Privacy policy, terms, and a security page. The security page explains the architecture and how to verify it yourself in DevTools, rather than asking you to take our word for it.
- Full API reference with five languages, every error code, and machine-readable copies at
/api.mdand/openapi.json. There is a one-click copy that hands the whole reference to an AI assistant. - Accounts are here. Sign in with just your email, no password, and get an API key plus a full history of every dedupe you run.
- A real dashboard. Usage over time, duplicate rate per run, rows kept versus removed, API latency, a monthly quota gauge, and a breakdown by endpoint. Every figure comes from your own runs; nothing is a sample.
- The API is metered in rows, not calls. One unit is one row of up to 1 KB, so a 50,000-row file costs the same whether you send it in one request or fifty. Free includes 1,000 rows a month, Starter 100,000, Pro 300,000.
- Automation guides and a comparisons section. A copy-paste Apps Script that adds a Dedupe menu to Google Sheets, a Zapier/Make/n8n guide with a downloadable n8n workflow, and five honest comparison pages that say when the other tool is the better pick.
- Four new guides: Google Sheets, Excel formulas, Android photos, and CRM contact cleanup, joining the iPhone, Google Photos and Mac guides under a single guides index.
- Export everything you have stored, as CSV or JSON, and delete your account outright. The server only ever holds job metadata: counts, timings and column names, never your file contents.
- Rebuilt the site on Next.js. Same addresses, same in-browser tools, nothing uploaded.
- Fixed a request-size limit we could not actually honour. Requests are capped at 5 MB now, which is what the platform genuinely serves, rather than failing at a larger advertised number.
- Fixed the mobile layout, where the header had been taking up nearly half the screen, and a download button that rendered invisibly against its own background.
- Redesigned the site around a dark theme.
- Pro waitlist submissions now land on a confirmation page instead of a 404.
- The homepage is generated from the same template as every other page, so its FAQ markup can no longer drift from the FAQ you actually see.
- Added
llms-full.txtalongsidellms.txt. - Spreadsheet parsing now loads only on pages where you can drop a spreadsheet, cutting about 400 KB from the text and JSON tools.
- Fixed CSV files saved from Excel as "CSV UTF-8" being read with a stray character in the first column name, and tab-separated files being split on commas.
Last updated