How we find emails that actually work

Finding someone's email on the public internet sounds simple until you try it at scale. The problem isn't finding emails — it's finding the RIGHT email. The one that belongs to this specific person, not a collaborator on a shared repo. The one that's personal, not a work inbox that'll bounce in three months.

PersonFinder looks at public GitHub commits and profiles, then portfolios and resumes. If that trail is thin, we search the open web with DuckDuckGo. There is no paid search API in the default path.

“GitHub commits, portfolios, resumes, then DuckDuckGo. MX checks the domain. A model picks from found addresses only — it cannot invent one.”

Commit history is often the richest source: the author email on a commit was set by the developer. We use the GitHub commits API, and when that is rate-limited we fall back to a commit-only git fetch.

Every candidate email is checked with DNS MX. That confirms the domain can accept mail. It is not SMTP proof that the mailbox exists, and we do not guess first.last@company.com.

A model then picks a likely personal email from that list. It may reject work inboxes and generic aliases. It must not invent an address that was never found.

Hit rate depends on how public the person is. We do not publish a fake 86% figure. Plans start at ₹99/month. MCP is on Scale.