Munich Coding Vibes — Windsurf × Aparavi Workshop Recap
We teamed up with Windsurf at Aparavi’s Munich office on 07/31/2025 for a fast, hands-on night of building production-ready, natural-language apps. Using Windsurf’s agentic IDE and Aparavi’s MCP server, attendees turned raw PDFs, videos, and spreadsheets into working AI apps, then battle-tested them in a friendly code challenge.
Workshop highlights
Aparavi intro — How MCP exposes governed data sources to LLMs for secure, prompt-centric workflows.
Windsurf walkthrough — Flow-state coding with an agentic IDE, CLI setup, and project scaffolding.
Live build — Connect Windsurf CLI to Aparavi’s MCP server, query real data, and ship an app in under an hour.
Code challenge — Small teams built features, pitched lightning-style, and took home prizes.
What attendees learned
-
How to wire Windsurf + Aparavi MCP for natural-language development on unstructured data.
-
How to orchestrate data access across PDFs, images, and sheets without copying data.
-
Practical patterns for chunking, retrieval, and tool calling in agent workflows.
-
How to move from prototype to repeatable app with CLI, prompts, and connectors.
Windsurf × Aparavi in practice
We demonstrated an end-to-end path: configure MCP, mount governed data, iterate prompts in Windsurf, and deploy a working agent that answers questions, cites sources, and respects data boundaries.
All the photos are available here: https://photos.app.goo.gl/2x6tS2izzazvWdce7
Resources and next steps
-
Try Windsurf: visit windsuf.com for downloads and docs.
-
Explore Aparavi’s Model Context Protocol server and data toolchain to build your own pipeline.
Photo gallery




