Connectors are hard. Until they’re not.
At Meltano, we believe data integration shouldn’t require a team of senior engineers. With Meltano’s open-source framework, Meltano Cloud’s intuitive GUI and the power of AI-assisted development through Claude Code, anyone can build a production-ready connector even on their first week on the job.
Here’s how one of our newest team members proved exactly that.
Starting from Zero: A New Employee’s First Week
When I joined Meltano, I came with enthusiasm but little else. No background in data pipelines, no knowledge of YAML and no idea what a connector actually was. My first task? Figure it out.
I started by logging into Meltano Cloud and working through the documentation to get set up locally. With hundreds of plugins available in the catalog, the sheer variety was initially overwhelming. But rather than being paralyzed by choice, I found my entry point: a simple, no-authentication-required plugin called Fuel Prices.
No API keys. No secrets. Just data: 1,202 rows of weekly UK fuel pricing from 2018 to present, visualized cleanly in DBeaver.
It was a small win, but it was mine.
Exploring the Ecosystem
With the basics under my belt, I started exploring further. Using Postman to inspect my workspaces and pipelines, I began to understand the broader architecture of what Meltano makes possible.
Then came the real challenge: building something from scratch.
Building a Custom Connector with Claude Code
Every good origin story needs a catalyst. Mine was a burning question: What group is my favourite dog breed in?
Enter DogAPI, my first custom connector.
I had no idea how to write a Meltano YAML file. I didn’t know where to start with plugin structure. What I did have was Claude Code, Meltano’s framework, and an API endpoint.
Here’s the process, simplified:
- I gave Claude Code the essentials: the API endpoint, the two call methods, and sample output files so it understood the data shape.
- I shared the documentation link and let Claude Code run in plan mode.
- Claude Code did the heavy lifting, editing the Meltano YAML file and scaffolding out the plugin folder structure automatically.
- When errors surfaced on the first test run, Claude Code identified and fixed them before I even had a chance to read the error message, with my approval of course.
The result?
A fully functional local connector, built with no prior experience in connector development.
From Local to Live: Deploying to Meltano Cloud
A connector that only runs on your laptop isn’t much use to anyone. Getting it into production was the next step, and Meltano Cloud made it straightforward.
Every Meltano Cloud workspace is natively linked to a GitHub repository, purpose-built for exactly this kind of workflow. Here’s how deployment went:
- Committed the local connector using Git
- Resolved a merge conflict in VS Code
- Pushed to GitHub
- Clicked Deploy in Meltano Cloud
Within minutes, the DogAPI connector was live. I added it to a pipeline alongside a Postgres Warehouse and dbt, and it was immediately integrated into my full Meltano workspace, no extra configuration needed.
What This Means for Your Team
Sure, DogAPI is a simple use case. But the principle scales.
The combination of Meltano’s pre-built, open-source framework, the visual simplicity of Meltano Cloud, and AI-assisted development through Claude Code has fundamentally changed what’s possible for teams without deep data engineering resources.
Going from zero experience to a deployed, pipeline-integrated connector in a couple of days isn’t a trick. It’s the product working as intended.
Whether you’re a solo analyst, a scrappy startup or a growing data team, Meltano removes the barriers that used to make custom connectors a specialist’s job.
Ready to Build Your Own?
Start with Meltano Cloud and see how far you can get, faster than you’d expect.