Feldera docs

Examples and guides for building incremental feature pipelines on streaming data.

Get started

Browse the docs

Connect sources, write incremental SQL, and serve model features.

Quickstart

Install Feldera, create an incremental SQL pipeline, and query features over HTTP in about ten minutes.

API Reference

Docs cover pipeline management, feature queries, connector setup, and authentication endpoints, with request and response schemas.

Connectors

Use Kafka topics, Postgres CDC replication slots, MySQL binlogs, and Amazon S3 buckets as live pipeline inputs.

Core ideas

Learn incremental SQL, operator graphs, delta state, and feature serving, plus how Feldera keeps results current.

Three steps

Start with your first pipeline

Run Feldera locally or in managed cloud. Choose a path and begin.

1

Install

Pull the Feldera Docker image or install the CLI. Self-host in minutes, no Kubernetes needed.

2

Write SQL

Write feature logic in standard SQL. Feldera compiles it into an incremental operator graph, updating with each new event.

3

Feature lookup

Features are available over HTTP. Call /features/{pipeline}/{key} from model serving or your feature store.

Go to Quickstart