- Core concepts — The ideas everything else rests on: everything is a document, the authoring pipeline, logic as data, access by scope, encryption under a…
- Offline-first and the sync model — Why the engine is embedded, how the stack stays reactive from one changes feed, what replication carries and why it bypasses the pipeline…
- Schema engine — How a class document becomes a validator, how attribute types map to Zod, and how a schema change propagates to the documents that already…
- Query engine — How a SQL statement becomes rows: parse, plan, execute; what the planner pushes down to the index and what it evaluates in memory; how…
- Crypto engine — The key hierarchy today, the canary, locked stacks, key identifiers and the keyring, which reads decrypt and which deliberately do not, and…
- Access control — DocStack has one access-control language, a monotone attribute formula sealing a scope of content, and it is enforced by cryptography…
- Transactions — Why the stage sits above the storage plugin and commits through it, how overlay reads and the validation sweep work, what atomicity…
- Patches & migrations — Why the data model ships as versioned patch documents, why a chain applies through one transaction, how deferral behind the document key…
- Security overview — The five properties the engine enforces on its own, how sessions are established, what leaves the device, and the limits stated plainly.
- Under the hood — The libraries the engine is built on, the engines and where each lives in the source, and the patterns that recur across them.