Admin documentation

What Teloma is

The product in one page — what it does, who it's for, and what it deliberately doesn't do.

Read →

Cancer, treated as an observability problem

The mental model behind the app — why this maps cleanly to logs, metrics, events, and outcomes.

Read →

Cancer-agnostic by design

Why fields are open — and what to do when your cancer type isn't on a drop-down.

Read →

Privacy and consent — the short version

How sharing works: by default nothing leaves you; everything else is opt-in and revocable.

Read →

Public Cancer Data Bank

A metadata-first registry of open oncology datasets — sources, datasets, drugs, combinations, trials, and ingestion jobs.

Read →

Audit log

Every view, write, delete, export, and DENY — append-only, indexed by actor and entity.

Read →

Research exports

Generate Safe-Harbor de-identified exports from consented patients — logged as EXPORT events.

Read →

What Teloma will not do

The hard product limits, written out plainly.

Read →

AI guardrails (Research Assistant)

Three layers of safety so the AI can never recommend, rank, or dose.

Read →

Architecture

Two services over Postgres + object storage. Where each concern lives.

Read →

Auth & RBAC

Roles, access grants, the resolve_patient dependency, and how DENY gets audited.

Read →

Data model

What the tables are, what they hold, and the few invariants you need to know.

Read →

Data Bank connectors

How the three live connectors talk to public services — and where their results land.

Read →

Runbook — local dev

Start everything, reseed, and run the AI host.

Read →