Manufacturer of lighting solutions, 2026

An MVP with AI integration for a lighting manufacturer.

piparo has been working since January 2026, via a placement partner, on a trade-fair workstream: effort estimation via work packages, PoC, visit reports and an LLM integration with MCP servers and Langfuse observability.

Screenshot to follow

Screens

This is how it looks.

The footage from this project hasn't been released yet. It's coming soon.

Screenshots to follow
Screenshots to follow
Screenshots to follow

Problem

The initial situation.

An industrial customer needed an MVP with AI integration on short notice, without building an internal LLM and MCP infrastructure for it.

Solution

How we solved it.

Piparo estimated the scope via work packages, delivered PoC and visit-reports work, and set up the AI runtime with a Rust-based genai fork, MCP servers and Langfuse observability.

Result

What came out of it.

The workstream has continued since January 2026. MCP expertise is now established at piparo as a standalone service.

Effect in figures

Which was measurable in the end.

01 January 2026 Ongoing workstream
02 A trade fair workstream MVP, PoC, Visit Reports Workstreams
03 Subcontractor via GTS placement Role

Information

How it was built.

Platforms
  • Backend
Tech Stack
  • Rust
  • Genai
  • MCP (rmcp, rmcp-sse)
  • Langfuse
Architecture
LLM runtime integration based on a genai fork in Rust, complemented by an MCP server ecosystem (rmcp, rmcp-sse, mock-llm-server) for tool and resource access, as well as Langfuse for AI observability.
Duration
Since January 2026

Next step

Do you want to build something similar?

30 minutes is usually enough to find out whether we are the right people for your project.