JP Marhefka
Skip to case study

02

Partuno

I built Partuno to compare DigiKey and Mouser components, then made the recommendation evidence visible.

Role
MCP and backend engineer
Year
2026
Discipline
Python, API design, model tooling
Partuno project poster
Scroll to case study

01 Overview

What I built

Partuno is an open-source MCP server for comparing electronics offers across DigiKey and Mouser.

Saved Partuno demo comparing a ten-piece component offer across DigiKey and Mouser
Saved demo of a 10-piece comparison: the same $1.71 parts total appears at both distributors; packaging, stock, and fulfillment can still change the choice.

How it works

The server turns distributor responses into normalized component records before comparison, risk review, and substitute evaluation.

That gives an engineer one place to see the part, the sourcing context, and the reason a recommendation was made.

The public repository gives other engineers a place to inspect the workflow and extend it.

02 Challenge

Where it got hard

The hard part was bringing vendor APIs, pricing context, and comparable recommendations into one workflow.

Saved Partuno demo showing a BOM risk overview
Saved public demo of BOM-risk review, where lifecycle and sourcing context are surfaced before a comparison is accepted.

The sourcing problem

Normalized metadata lets one request compare vendors on consistent attributes instead of stitching together raw responses.

Recommendations keep pricing, availability, lifecycle, and substitution context beside the suggested part.

Credentials stay with the operator through local and self-hosted MCP deployment paths.

03 Process

How I structured it

I separated distributor adapters from the records and tools that support a sourcing decision.

Saved Partuno demo showing product research results
Saved demo of product research across normalized distributor records.

The implementation

I integrated distributor APIs and normalized part attributes into schemas the comparison logic can use.

I added pricing visibility, lifecycle and compliance checks, lead-time context, and substitute discovery support.

I added local credential paths, optional remote or self-hosted deployment modes, and packaging for MCP clients.

Technical illustration of Partuno’s distributor normalization and comparison workflow
How distributor responses become normalized records for comparison, risk review, and substitute evaluation.

04 Outcome

Where it landed

Partuno is a public tool for component research, comparison, and substitute discovery.

Saved Partuno demo showing recommendation evidence for a component choice
Saved public demo of recommendation evidence: the workflow keeps the reason for a suggested component visible alongside the result.

The result

The workflow gives me a repeatable way to compare parts and find substitutes across distributor sources.

The public repository makes the comparison workflow inspectable and gives other engineers a starting point for extending it.

Status
Public repository
Public project
Partuno on GitHub
Primary purpose
Component sourcing and BOM comparison