Building with Redpanda Connect: Bloblang and Claude plugin

Redpanda Connect is built to make streaming pipelines feel straightforward with 300+ connectors, a declarative config model, and Bloblang for transformations. No JVM, no extra infrastructure, no ceremony. In this workshop, you'll build a production-ready Postgres-to-Redpanda pipeline and see what that simplicity looks like in practice, including how the Claude plugin accelerates the whole authoring workflow.

We'll use a concrete, practical example throughout: capturing changes from a Postgres database, transforming the payload with Bloblang, and streaming it to Redpanda. By the end, you'll understand not just how to wire things together, but why the architecture works the way it does.

What you’ll learn 

Here’s what you’ll get a deeper dive into:

Redpanda Connect architecture. The three-component model (inputs, processors, and outputs) is the foundation for every pipeline you'll ever build. Everything we build in this workshop starts here, so we'll take the time to make it feel intuitive.

CDC from Postgres. We'll use the postgres_cdc input to capture row-level changes from PostgreSQL's Write-Ahead Log (WAL),

Bloblang transformations. Redpanda Connect's native mapping language lets you filter, reshape, enrich, and route data, all without leaving your pipeline config. You'll write real mappings against real data.

The Claude plugin. The Redpanda Connect Claude plugin runs inside Claude Code and understands the full Connect spec, including inputs, processors, outputs, Bloblang syntax, and common configuration patterns. You'll use it to generate pipeline configs, explain transformations you didn't write, and debug errors on the fly.

Testing and iteration. A pipeline that works on your laptop is the prerequisite for a pipeline that works in production. We'll cover local validation and the debugging workflow that actually saves time.

How the workshop is structured

Part 1: How Redpanda Connect works 

(10 minutes)

Before we build anything, we'll take a few minutes to walk through how Redpanda Connect actually works. Inputs, processors, and outputs: three components, one clean config, and a surprising amount of power. If you're new to Connect, this will give you the mental model you need to follow along with everything that comes after.

Part 2: Building your pipeline with Claude 

(25 minutes)

With your environment set up following the prerequisites, we'll jump straight into building. We'll use the Claude Plugin to generate and refine a complete Postgres-to-Redpanda pipeline together, starting from nothing but a plain-language description of what we need. By the end of this section, you'll have a working pipeline and a solid feel for how the Claude Plugin fits into your development workflow.

Part 3: Advanced Bloblang and enrichment patterns 

(20 minutes)

This is where things get fun. Once the base pipeline is humming, we'll extend it with conditional routing, schema enrichment, and error-handling patterns that show up in almost every real-world pipeline. If you've dabbled with Connect before and want to level up your Bloblang skills, this is the section for you.

Part 4: Best practices and Q&A 

(15 minutes)

We'll wrap up with the production considerations that actually matter: config validation and observability.

What you’ll need

You’ll have two Redpanda experts walking you through the whole process, but it’ll certainly go smoother if you already have the following:

• Basic familiarity with data streaming concepts
• YAML literacy (helpful but not required, you'll pick it up)
• A laptop with Redpanda Connect installed for local testing
• rpk installed
• ‍Claude Code subscription: The Redpanda Connect Claude Plugin runs inside Claude Code and uses your existing plan's token allocation. No subscription? No problem. You can still follow everything in the workshop. You'll just build and debug pipeline configurations manually rather than with the plugin.

Start building your pipeline

Whether you've never written a Connect config or you're looking to sharpen your Bloblang skills and explore AI-assisted pipeline development, this workshop is built around real problems and working code. Here’s what you’ll leave with:

• A working Postgres CDC pipeline streaming to Redpanda
• Hands-on Bloblang experience you can apply to any transformation challenge
• A practical Claude plugin workflow you can drop into your next project immediately
• The pattern knowledge to build pipelines your team can actually maintain

Ready to start building? Join our hands-on workshop and we’ll see you there!

In the meantime, browse the Redpanda Connect documentation, explore the 300+ available connectors, or sign up for a free trial to spin up a cluster before the workshop.