Drowning in manual entry of old characters, strike-throughs, and re-created registers?
"I can't read the old characters in the family register; entry takes hours…" "There's no system to track ancestry-research progress; we get by with notebooks and Excel…" "Paternal and maternal lines get tangled, the management gets confusing, and organizing barely moves forward…"
To solve these on-the-ground problems at legal scrivener offices specializing in ancestry research and inheritance, Nexa Engineering built our own AI family-tree SaaS, "Nezasu by Nexa."
Overview
Nezasu by Nexa is a cloud SaaS that fully digitizes the ancestry-research work scriveners perform.
- AI-OCR auto-reading — Claude Vision API recognizes old/variant characters and strike-throughs with high accuracy. It supports multiple register formats and extracts name, birth date, relationship, and registered domicile in a structured, cell-level way.
- 4-level case management — Organized as case → ancestral line → person. A work checklist, expense details, and order-date management are consolidated on one screen.
- Line color identification — Set an identifying color per line (paternal, maternal, etc.) so "whose ancestor is this" is instantly clear.
- Auto family-tree generation (in development) — Generate a family tree in real time from registered people and relationships. Interactive with zoom, pan, and node editing.
- Register-acquisition guide (in development) — Automatically suggests the next municipality to request from based on prior-register info, preventing gaps.
- Notion / PDF export — One-click copy of all case info in Markdown — the base of a client report is ready.
Tech stack
| Area | Technology |
|---|---|
| Frontend | Next.js (App Router) / TypeScript / Tailwind CSS |
| Backend | Supabase (PostgreSQL + Auth + Storage + Realtime) |
| AI | Claude Vision API (Anthropic) |
| Multi-tenancy | Full isolation via Row Level Security + JWT app_metadata |
| Infrastructure | Vercel + Supabase Cloud |
Nezasu, too, was born by changing the "order of building"
Nezasu didn't aim for the finished form from day one. We first released with only the minimum project-management and AI-OCR features, then heard how scriveners actually used it on the ground. Watching real usage, we built out the features that truly mattered, one by one.
This let users start early, and prevented the "this isn't what we wanted" trap that comes from building everything up front. We practice the very approach we propose to clients — "a working thing first, before requirements" — on our own product.
Try it first
"I'd like to test whether it fits our work" or "Could it apply beyond ancestry research?" — please reach out.