How We Built a Living Family Tree for 170+ Relatives — Without Losing the Story

View the live project at babafeqi.raafat.site, then contact RAAFAT-Technologies if your family, community, or organization needs a private heritage platform — we'll scope a V1 around your actual workflow, not a template.

The problem isn't "no family tree apps" — it's trust and context. Commercial tools assume Western naming, public profiles, and subscription models. This family needed Harari heritage preserved, duplicate-name disambiguation (many "Ahmed"s and "Yonis"s), and admin review before anything goes public.

Relational data beats nested JSON for real family trees. People, parent–child links, and marriages live in separate PostgreSQL tables so branches can grow independently, relationships stay correct, and the tree scales from ~170 seeded relatives toward thousands without rewriting the frontend.

Three experiences, one source of truth. Public visitors browse and search without an account. Registered relatives submit or update their record through a moderated "Add yourself" flow. Record keepers use an admin dashboard — tree editor, submission queue, backup/restore — so corrections don't require a developer.

Cultural details are product features, not polish. Ethiopian calendar dual-date fields, ancestry-path search results, patronymic display rules, and a homepage that tells Faqih Yonis Abdosh's story turn a database into something the family actually recognizes as theirs.

Ship a focused V1, then let the record grow. Interactive tree, search, relationship finder, statistics, and backup shipped in ~3 weeks. Social feeds, AI chat for visitors, and photo archives can wait — the core loop (find a relative → see how you connect → submit a correction) had to work first.