Command Context Consistency
Commands declare the event subset they depend on with JSON criteria, then save only if that context is still current.
Per-boundary orderEvents publish in ascending log position.
Durable checkpointsPublisher state is stored with the selected backend.
Same APIsSQLite and PostgreSQL expose the same gRPC surface.
Why Orisun
The event log, consistency boundary, and delivery loop live together.
PostgreSQL or SQLite is the durable source of truth. Embedded JetStream is the delivery layer. Your application keeps business decisions in application code.
No skipped committed events
Publisher checkpoints live in PostgreSQL or SQLite. Wake-up signals can be missed; committed events still drain sequentially.
One deployable server
Event storage, indexed reads, auth, the gRPC and Admin APIs, embedded NATS JetStream, and telemetry ship in a single binary.
Where it fits
Use Orisun when correctness depends on the event log, not just message delivery.
Orisun is for systems where commands need to read event history, make a business decision, commit exactly the events for that decision, and publish the committed log in order.
Embedded event store
Link Orisun into a Go service when event storage should live inside the same process and deployment unit.
Standalone event service
Run Orisun as a small gRPC service when multiple applications or languages need the same event API.
Reliable projectors
Build read models from catch-up subscriptions that recover from downtime without depending only on broker retention.
Backend choices
Use the smallest operational shape that fits.
SQLite is complete for single-node systems. PostgreSQL adds multi-node Orisun coordination and database-managed operations without changing client code.
SQLite
Open guideSingle-node production, embedded apps, local development, and low-ops services.
- event log
- admin state
- indexes
- publisher checkpoints
PostgreSQL
Open guideMulti-node Orisun deployments backed by database-managed storage and operational tooling.
- cluster locks
- schemas
- PgBouncer-friendly config
- shared storage
Command flow
From decision to projection, the ordering contract stays explicit.
Orisun keeps storage, consistency checks, publishing, and subscriber recovery on one ordered path so application code can focus on domain decisions.
Declare context
Select the event subset your command depends on with JSON criteria.
Save atomically
Commit only if the selected context is still at the expected position.
Publish in order
Drain committed events from durable checkpoints into embedded JetStream.
Project safely
Catch up from storage, then consume live events with idempotent checkpoints.
Documentation
Find the right page for the job.
The docs are organized around the work teams actually do: start a server, model a consistency boundary, integrate the API, embed Orisun, and operate it.
Start
OverviewUnderstand where Orisun fits and which path to take.
Getting StartedRun SQLite or PostgreSQL as a binary, container, or embedded store.
TutorialBuild a ledger with CCC, indexes, and a live projector.
Understand
ConsistencyModel business invariants over event-content queries.
Delivery guaranteesSee how checkpoints preserve ordered publishing.
Storage backendsChoose the right backend for your deployment shape.
Build
EventStore APISave, query, subscribe, and manage indexes over gRPC.
Go embeddingRun Orisun inside a Go service with one backend package.
OperationsDeploy binaries or containers, then configure auth, TLS, storage, and telemetry.
Get moving