We hateDuplicates.
Every contact database rots the same way. Imports pile up, names fork into three spellings, and six months later nobody trusts the CRM — so deals get double-called, reports lie, and someone spends a Friday with a spreadsheet and a headache.
Epafi exists to end that. Not with another place to store contacts, but with the infrastructure that keeps them clean: match, import, enrich, relate. One record per person within the account, with every merge remaining reviewable. That is the whole product.
Four things werefuse to bend on.
Names are messy
People spell names differently, change jobs, use two phones and three emails. Any tool that pretends otherwise is lying to you. Epafi is built for the mess as it actually exists.
Merging is a decision
A merge destroys information if it is wrong. That is why every match carries a confidence score, automatic merges use a threshold you set, and each merge stores the snapshot needed to undo it.
Your data is yours
Your contacts are not our product. We do not sell data, we do not train shared models on your records, and everything exports cleanly — CSV, JSON, or API — whenever you want.
Speed is a feature
Duplicate review has to fit into normal contact work. Epafi supports explicit review, threshold-based matching, and scheduled workflows without hiding the decisions that changed a record.
3
Applications
Web + REST
Interfaces
Clerk
Authentication
Polar
Billing
Small team.Sharp elbows.
We are a compact, distributed crew of data engineers and designers across Athens, Berlin, and remote. No growth team, no sales floor — the product is the pitch.
No dark patterns, ever
No "AI-powered" as a substitute for an answer
No growth hacks in the product surface
Docs written by the people who built the API
Support answered by engineers