ParishBrain is not an app you install. It is the whole stack — managed, on–premises, and run for you.
ParishBrain is AI–native and turnkey. We bring the servers, run the service, and maintain the software — no programming, no app development, no setup on your end.
We provision it, run it, update it, and keep it healthy — remotely and on a schedule you control. You do not hire an IT department to use ParishBrain; that is our job.
On–premises boxes we bring to your campus and hook up. Your data and your AI live here, inside the building — never rented from someone else’s cloud.
One connected suite of apps for church and school, each replacing a subscription you already pay for, all reading from a single encrypted identity layer.
Which means an institution can cancel the subscriptions it stitches together today.
Church and school apps sit above shared operations. Everything reads from and writes to FamilyGraph — which clarifies who a family is inside the building and anonymizes them everywhere else. The flagship pieces are marked in gold.
Data flows in from every source. Giving, enrollment, communications, transcripts, sacramental records, volunteer hours. Every system that touches a family contributes.
FamilyGraph resolves identity — and anonymizes it. The encrypted core deduplicates across systems into one coherent profile per family, then issues the pseudonyms that let cloud–facing apps work without ever seeing who that family is.
Context accumulates over time. A giving pattern across five years, an engagement trajectory across three grades. The graph grows richer with every record.
A local model activates it. An on–premises model reads the accumulated context to surface what no administrator could see by hand — nothing sent to a frontier provider.
Decisions improve at every level. The pastor knows which families to call; the principal, which teachers need support; the development director, which alumni to steward.
No data leaks. All PII and model inference run on–premises. Pseudonyms serve the cloud apps; real identities never leave the building.
No frontier dependence. A capable model runs on your own server — nothing sent to OpenAI, Anthropic, or Google. You own the weights, and it keeps working with the internet down.
No misuse, by anyone. Role–based guardrails for every stakeholder: a Socratic mode and no–write guard for students, sensitive–content alerts, and an immutable audit log.
We’re building with a small number of Catholic institutions. Leave your email and we’ll be in touch.