MCP launched: 176 projects onboarded, 69 tools enhanced, and AI-assisted actions grew by 630%. Teams now describe what they need, and AI builds pipelines, writes SQL, and configures components.
When we officially launched our Model Context Protocol (MCP) server integration on June 12, 2025, we weren't just adding another feature - we were fundamentally changing how data engineers interact with their tools. One month later, the transformation has exceeded our wildest expectations.
The Model Context Protocol represents a paradigm shift in data engineering. Instead of clicking through complex UIs or memorizing syntax, engineers can now describe their intentions in plain English. Need to extract data from a REST API? Simply tell the MCP-enabled extractor what you're looking for. Want to write a complex SQL or Python transformation? Describe your business logic and watch as intelligent code generation handles the SQL.
MCP transforms every Keboola component into an AI-powered assistant that understands context, suggests optimizations, and can even debug issues. It's not just automation - it's augmentation of human expertise.
In our first full month since launch, the MCP server integration has achieved remarkable traction:
The month-long adoption reveals how transformative MCP really is:
Generic extractors lead adoption with 527 AI interactions across 37 projects.
The Real Game Changer: Accelerating Development Velocity
A full month of data reveals something profound about how MCP changes development:
This isn't just adoption—it's transformation. Engineers are using MCP not just to build faster, but to build better, with AI helping them explore possibilities they might not have considered. And this is just the beginning.
A month of real-world usage has taught us valuable lessons:
Setup Complexity Evolution: Initial setup friction drove us to develop our streamlined MCP Quick Start flow, reducing configuration time significantly, enabling the Oauth authorization. The data shows that teams with proper setup achieve 3x higher engagement rates.
Scale and Performance: Month-one usage patterns revealed optimization opportunities. We've implemented improvements that handle peak loads of 300+ daily interactions seamlessly.
These challenges weren't roadblocks - they became our product roadmap. Each friction point revealed an opportunity to make conversational data engineering more accessible.
A month of usage revealed some surprising patterns:
Generic Extractor Dominance: While we expected transformation components to lead, generic extractors (527 actions) showed the highest single-component adoption, indicating teams are using MCP for rapid API integration.
Recent Acceleration: The last week shows 523 actions across all projects, indicating sustained momentum beyond initial experimentation.
What we're witnessing goes beyond impressive adoption metrics. MCP represents the emergence of conversational data engineering - where the barrier between human intent and technical implementation dissolves.
The first month data shows:
With 523 MCP interactions in the past week alone, we're seeing clear evidence that conversational AI has become essential to modern data engineering workflows.
We're not just automating tasks; we're augmenting human capability at every layer of the data stack - and the first month shows this augmentation is becoming indispensable.
Ready to experience conversational data engineering? Create a free project and try on your own.
