Your AI agent just became a data engineer. Keboola MCP Server lets Claude, Cursor or ChatGPT build actual data pipelines—with error handling, logging & built-in governance. Describe your workflow and get production infrastructure in seconds.
Free Options
Launch Team
Render: Get up to $100K in startup creditsThe fastest way to deploy and scale apps and agents
Promoted
Lovable for Data Pipelines: Revolution in Data Engineering
Hi all, I am Pavel one of the co-founders. Ever needed to get data from multiple sources, put it together and send it to your Salesforce or Google Ads ? Remember how difficult it was and how many hours your team had to spend on it ?
Starting today you can vibe code data pipelines yourself and get the data when you need it! Together with Claude or Cursor this becomes a reality.
Check it out and let me know what you think :)
Here is video in action by Avery:
@pavel_dolezal2 :) Proud to be part of this!
Say hello to the future of data engineering :)
I'm not going to play it cool, because this is a big one ☺️
We built this for every data engineer who’s ever stared at a failed job, tired of chasing nulls, fixing broken DAGs, a silent timeout, or for every “ran successfully, but returned 0 rows” mystery. “Seriously? I just wanted to join two tables.”
With the Keboola MCP Server, you just describe what you need.
And your AI agent (Claude, Cursor, ChatGPT) builds a real, production-grade pipeline for you.
Versioned. Logged. Governed. Actually usable. And fast.
⚙️ Try it if you’re curious.
🙏 Vote if it resonates.
🚀 And may your pipelines run smoothly.
We’re proud of what we built. And I'm even more excited to see what you build with it 🩵.
🔨 Ready to Start?
Full docs + usage examples on GitHub
👉 https://github.com/keboola/mcp-server
Product info, use cases, and demo:
👉 https://www.keboola.com/mcp?utm_campa...
Hi everyone! I saw so much support today :) Thank you.
Our #KeboolaMCPServer is for every data engineer or data analyst who’s ready for something better. We built this for you!
The era of manual pipelines is over.
At Keboola, we’ve made it our mission to change how Data Engineering works forever. Back when we started building everything on top of APIs, we didn’t know it, but we were building for the agentic era.
And it is here.
We believe pipelines shouldn’t take weeks to build, require five tools to maintain, or wake you up at 3 a.m. to fix.
And with Keboola MCP Server, your can now:
➡️ Talk to your pipeline, and it builds itself.
➡️ AI agents like Claude or Cursor don’t just suggest code, they deploy real, production-grade workflows.
➡️ Start shipping faster, safer, smarter.
🩵 If you want to get started with right now, here!
Even though i understand keboola very well, this saves me time and sometimes it's smarter than me. It's ridiculous how clever the LLMs got. I'm mostly using it in web claude, but have it in cursor too.
I love how i can give it task and just go about other work for a few minutes.
@tomas_fejfar Totally agree. It’s been surprising to see just how capable LLMs have become, especially when they’re connected to real infrastructure like this.
That moment when you can hand off a task, shift focus, and come back to something working… it really changes how you think about building :)
Thanks for sharing and we are glad to hear it’s saving you time. Curious to see how you keep using it. 🩵
A couple of questions which I have:
1 - Does Keboola have its own compute and storage on which it runs the data pipelines or can it deploy pipelines to any platform such as snowflake, databricks, AWS etc
2 - Can this also be used to deploy infrastructure on cloud platforms?
Good question@daniyal_ahsan.
1. Yes, Keboola runs data pipelines on its own managed infrastructure, but it also integrates tightly with external platforms such as Snowflake, Databricks, BigQuery, Redshift, and more. Keboola uses its own compute and storage layer to orchestrate and run pipeline jobs. However, you can leverage external compute engines for transformations, like Snowflake, dtb, BigQuery etc.
2. Keboola is not an infrastructure provisioning tool. While it runs data pipelines, integrates with cloud platforms, or allows using external compute, It is not an infrastructure-as-code or cloud provisioning tool.
@dessignnet Thank you! :)
Really curious if this could let non-engineers describe what they want and get something usable. Any plans to add natural language prompts specifically for marketing use cases?
Hi @hamza_afzal_butt !
This is exactly where we’re headed. We have a list of use cases with system prompts that guide LLMs—helping them steer the discussion by identifying root issues, suggesting solutions, and even “signing off” PRD documents. We’re building the entire pipeline for you, so it’s not just for data engineers anymore. This will also empower business teams.
Would you like me to follow up with you early next week? I should have a functional version ready by then.
This is an export from the long-term memory of my R&D model (see full list in Google Doc):
https://docs.google.com/document/d/14EI8oh32YbNIV0SFX6DLqtGM593leD2bsFxBwgSSkBI/edit?usp=sharing