Trax.Samples

Part of Trax: business logic you can call, schedule, or serve as an API, with every run recorded in your Postgres. Docs · Getting started · All repos

Trax.Samples holds complete Trax sample apps and the dotnet new templates for an API, a scheduler or a hub. It sits last in the stack and uses every other layer; the templates ship as Trax.Samples.Templates, and trax generate in Trax.Cli builds on the hub template.

Example

dotnet new install Trax.Samples.Templates
dotnet new trax-hub -n MyApp      # or trax-api, trax-scheduler
cd MyApp && dotnet run

Each template stores runs in memory, so it needs no database; swap UseInMemory() for UsePostgres(...) to keep them.

Samples

Each sample is a working app with its own projects under samples/. ChatService is the most complete.

Sample	What it shows	Storage
ChatService	A chat app: mutation trains, and a lifecycle hook that publishes each result to room-scoped GraphQL subscriptions over WebSockets, with a React client	SQLite
GameServer	A GraphQL API (API keys and JWT) that queues work, and a scheduler whose local workers run leaderboard, rewards and match trains, with the dashboard	Postgres
Bookworm	Two domains, each with its own schema and DbContext, joined by a cross-schema GraphQL edge; also adopts the architecture-guard fixtures	Postgres
GraphQLClient	A gateway calling two downstream Trax servers through keyed Trax GraphQL clients, all started by one dotnet run	In memory
EnergyHub	A hub (GraphQL, scheduler, dashboard) that runs no trains itself, and a separate worker process that claims and runs them	Postgres, RabbitMQ
ContentShield	An API that sends queued and run mutations over HTTP to a runner process standing in for a Lambda function	Postgres, RabbitMQ
PersistedOperations	An API that accepts only persisted operations, and a client that uploads, queries and hot-fixes them	Postgres
StateMachine	Two snapshot state machines (turnstile, checkout) served over GraphQL, with a React client driving them through the stateMachine mutations	Postgres
SignalRDashboard	A Blazor Server page showing live train lifecycle events from the SignalR broadcaster	Postgres
ApiAudit	Every GraphQL request captured as an audit entry and written in batches to a console sink	SQLite
JobHunt	The baseline of a job-hunt CRM: GraphQL, API keys, the dashboard and domain migrations, with a React client	Postgres
TestRunner	NUnit test projects run as queued trains, with results streamed to a React client over subscriptions	Postgres
DataPipeline	The scheduler running three Flowthru data pipelines as a dependency chain	Postgres

Where this fits

Trax is split into layers, one repo each. Take the ones you need; the trains you wrote do not change. You are here: Trax.Samples.

Repo	What it adds
Trax.Core	Trains, junctions and the chain, with no database and no DI container
Trax.Effect	A recorded run for every execution (Postgres, SQLite or in memory), DI, effect providers, the state-machine engine
Trax.Mediator	The train bus: run a train by handing over its input, with every chain checked at startup
Trax.Scheduler	Cron and interval schedules, retries, dead letters, and workers on other machines or in Lambda
Trax.Api	GraphQL generated from your trains, with authentication, audit and typed clients
Trax.Dashboard	A Blazor Server UI for runs, schedules and dead letters, mounted in your app
Trax.Cli	The trax tool: scaffold a hub and trains from an OpenAPI or GraphQL schema, and state-machine codegen
Trax.Samples	Complete sample apps, and the trax-api, trax-scheduler and trax-hub templates

Docs live in Trax.Docs and are published at traxsharp.net/docs.

Contributing

Read AGENTS.md before changing code. Report vulnerabilities privately as described in SECURITY.md.

License

MIT. There is no commercial edition, and there will not be one.

Trax is an independent open-source project and is not affiliated with the Utah Transit Authority, Trax Retail, or any other organization using the Trax name.