Fluxtion project starter

From a 10-line prototype to a native server

The Fluxtion starter and playground. No-friction development, right in your browser.

Playground examples

New to Fluxtion? Read the developer guide.

Fluxtion starter journeyDevelop and run in the Starter, Fluxtion, the Playground and your IDE. Fluxtion is the logical engine — deterministic, replayable, audited. Deploy it three ways: embedded in a runtime (Apache Flink, AWS Lambda, your application), standalone (run on the JVM or natively compiled with no JVM), or hosted inside the multi-tenant Mongoose server with its plugin ecosystem — audit logging, replay, web console, Kafka, Chronicle, FIX, Aeron, OpenTelemetry.FROM DEVELOPMENT TO PRODUCTIONDEVELOP & RUNStartershape the graphFluxtion cloudgenerate the processorPlaygroundrun in the browserIDErun · debug · replayFluxtionthe logical enginedeterministic · replayable · auditableembeddedstandalonehost in MongooseEmbeddeddrop into your processApache FlinkAWS LambdaYour applicationStandalonea self-contained appRun on JVMNatively compiled — no JVMMongooseindustrial serverMulti-tenant hostingAgent threads · adminPLUGINS — ENTERPRISE CAPABILITIESaudit loggingreplayweb consoleKafkaChronicleFIXAeronOpenTelemetryOne graph model.Author once — deploy to any target,deterministic and replayable.
Pick a starting point, each is a runnable project

One graph, four stops on the path. You begin embedded and interpreted, running straight in the browser, which is keyless. From there you grow into real I/O, host the graph inside Mongoose, then deploy it as a native binary. Everything below is a project you can run.

Prototype

Embedded

Fluxtion in your app

The graph runs inside your app. You feed events straight in and it reacts, with no server and no infrastructure. The quickest way to get a feel for the model.

Generation
Key every run Playground
Integrate

DataFlow connector

Feeds → graph → sinks

Wire up real inputs and outputs. File or in-memory feeds flow into the graph and named sinks take the results, all run on an agent thread. A proper multi-input, multi-output graph.

Generation
Key once at build Playground
Host

Mongoose + Fluxtion

The same graph, hosted

Run the same graph inside a Mongoose server. Connector plugins handle the enterprise integration for Kafka, files and the rest, with agent threads and an admin web console. One artefact, two runtimes.

Generation
Key once at build Playground
Deploy

Native image

Zero-overhead start-up

Compile the AOT graph to a GraalVM native image. It starts in milliseconds, with no JVM warm-up and no compiler at runtime.

Generation
Key once at build Playground
Templates are shareable and discoverable

Start from a community example or publish your own from any GitHub gallery. Every template opens here as a real, runnable project.

Browse all templates
Mongoose plugin ecosystem

Connect a hosted graph to the enterprise with ready-made plugins for Kafka, files and more. Browse the catalogue for the industrial server tier.

Explore plugins
The Fluxtion cloud service & API keys

The cloud service is a hosted source generator. For a downloaded project, the builder sends your graph to the service and gets back the dispatcher or optimised event processor, so you never install a generator toolchain yourself.

When do I need a key?
Playground — always keyless, build and run anything for free.
AOT download — a key once, at build time.
Interpreted / in-process download — a key on every run.
Get an API key on RapidAPI

Get Fluxtion

  • Download a starter project (.zip, runnable Maven project) · other templates
  • Repository: Current runtime, builder-api, builder, BOM and starter releases are published to the Fluxtion public repository, https://repo.repsy.io/mvn/fluxtion/fluxtion-public - declare it. Maven Central has only an outdated 1.0.9 runtime and builder and no BOM; the fluxtion-maven-plugin is published to Maven Central.
  • BOM: com.telamin.fluxtion:fluxtion-bom:1.0.82
  • Builder (build time): com.telamin.fluxtion:fluxtion-builder:1.0.82
  • Runtime: com.telamin.fluxtion:fluxtion-runtime:1.1.3
  • Starter (local authoring tool): com.telamin.fluxtion:fluxtion-starter-core:1.0.82
  • JBang header (one directive per line): 
    //REPOS mavencentral,fluxtion=https://repo.repsy.io/mvn/fluxtion/fluxtion-public
//DEPS com.telamin.fluxtion:fluxtion-builder:1.0.82
  • Remote build: Building a graph (DataFlowBuilder...build(), interpret(), Fluxtion.compile* / compileAot or the fluxtion-maven-plugin) sends the graph model - not your source files - to the hosted Fluxtion generator by default, and needs an API key. Without a key the build stops before sending anything ("Fluxtion API key is not configured"); a rejected key fails with HTTP 401/403. Put the key in ~/.fluxtion/fluxtion.apiKeyFile as apiKey=YOUR_KEY (the first build creates the file). A generated processor runs with fluxtion-runtime alone and makes no network call. How to get an API key — subscribe to "Fluxtion Source Generator" on RapidAPI, https://rapidapi.com/hub.
  • Node configuration is sent: The model includes the configured values of node instance fields, private fields included, rendered as Java expressions. Static, transient, synthetic and author-excluded fields are not captured, but that is not a secrecy guarantee: a value can still reach the model another way, for example through constructor arguments or custom serialization. Never put credentials in node fields: supply them at runtime with registerService(new Service<>(...)) and a @ServiceRegistered method, or with injectInstance(...). What is and is not sent: https://github.com/telaminai/fluxtion#cloud-generation-and-your-api-key
  • For agents and LLMs: /llms.txt · /CLAUDE.md