Runs in your infrastructure. All of it.
You deploy TypedFlows inside your own network, where it connects directly to your Kafka clusters and databases with exactly the permissions your administrator grants. Your data never crosses your boundary.
✓ Self-hosted ✓ Your network, your permissions
One application, next to your data
TypedFlows is an integrated solution: development environment, management console and the processing engine that runs your deployments. It keeps its state in a PostgreSQL database you provide, and talks straight to your infrastructure.
YOUR INFRASTRUCTURE · your network boundary
access
→
Sources & sinks · per environment
JDBC · sources & sinks
Where Kafka needs them
From design to a running deployment
Design
Compose the flow on the canvas; Java steps compile server-side as you type.
Validate
The design is checked end-to-end and step schemas are generated.
Deploy
Bind the design to an environment; override per-step values where needed.
Run
Start, pause, resume or stop the deployment. The engine does the rest.
Inspect
Every execution is recorded with its logs, right next to the deployment.
One instance, yours alone
TypedFlows is an integrated solution. It runs in your own infrastructure (a server, a VM or a Kubernetes cluster), packaged and rolled out the way the organisation already ships software.
A cloud provider is optional. If your Kafka clusters and databases already live with one, TypedFlows runs there too, inside your own account and under your own network rules. If you would rather not operate it yourself, we can run a dedicated instance for you. Whichever route you take, the instance is yours alone. There is no shared tenancy and no multi-customer control plane.
- Run it yourself and it stays inside your existing audit scope: your identity provider, certificates, backups and monitoring
- Run it yourself and it runs on capacity you already own
- Configured by your administrator
MINIMAL SETUP · illustrative
Kafka clusters, registries and database connections are added later, per environment, in the UI.
What the platform offers
What you actually work with, from the canvas to a running deployment.
Schemas validated as you design
Every step declares the data it expects. The platform checks the whole flow while you build it, before anything is deployed.
Clearly visible flows
The whole pipeline is one picture: every step, every branch, and the data that travels between them. You can see what a flow does without reading code.
Statically typed transformations
Write transformations in Java against typed input and output records. The compiler catches mistakes while you are still designing the flow.
Central deployment management
All deployments across all environments in one place, with their current status. Every run is recorded with its timings and logs, so you can open an execution and check its state.
The whole job happens in the browser
Canvas, step dialogs, deployment controls and execution logs all live in the browser, including a full code editor with syntax highlighting and inline compiler errors for Java steps.
Same design, every environment
A validated flow is promoted between environments unchanged. Connections come from the environment, not the design.
Exception handling you decide
The workflow declares what happens when a record fails: retry it, route it aside, or stop the deployment. Failures are recorded together with the record that caused them, so nothing fails quietly.
Schemas inferred from real data
Sample a Kafka topic, a database table or an HTTP endpoint and TypedFlows derives the schema from what is actually there.