Concepts
The domain model: holdings, accounts, vaults, groups, tokens, transactions, observations, the price graph, and the rollup. Every page ends with a why it’s modelled this way note. See Mental model.
Scani
One ledger across every asset you care about — exchanges, on-chain wallets, brokerages, banks, and manual entries. Same TypeScript codebase runs three ways.
What you’ll find here
Section titled “What you’ll find here”These docs are written for three audiences at once — operators self-hosting the stack, contributors reading the code, and agents ingesting Scani as context. Same content, organised so each can skim it.
Design decisions
Why the ledger is append-only. Why there’s no USD canonicalisation. Why transfer linking runs nightly. The cross-cutting choices that explain the shape of the code. See Design decisions.
Self-hosting
Three tiers from one binary. Compose for a one-box deploy, managed Postgres / Redis / S3 when you outgrow it, backup & restore, upgrades, observability, troubleshooting. See Tier model.
Reference
A full financial glossary — Scani-specific terms plus general finance. Database schema, tRPC routes, jobs, provider matrix, env-var index, and an LLM-friendly index.
New here
Section titled “New here” What is Scani What it tracks, what it doesn't, and the shape of the product.
Quickstart A self-contained stack in one command — no external API keys required.
How to read these docs An audience map: which pages matter if you're an operator, a contributor, or an LLM.
License & telemetry
Section titled “License & telemetry”MIT licensed. The OSS distribution sends no telemetry, ever — no install ID, no anonymous usage counter, no version-check beacon. The only outbound calls a self-hosted stack makes are the ones you explicitly configure.