Labsco
twominutereports logo

Partglyph MCP Gateway

from twominutereports

Deterministic industrial replacement engine with official catalogs and expert accuracy checks.

🔥🔥🔥Paid serviceNeeds API keys
#automation

Partglyph MCP Gateway

Partglyph MCP Gateway connects AI coding assistants and agents to Partglyph’s industrial replacement engine.

Use it from MCP-capable clients such as Codex, Claude Code, Cursor, VS Code, and other agent workflows to request replacement options, technical comparisons, source-backed matches, and previous match history without leaving the tools your team already uses.

Partglyph is built for deterministic industrial matching. It uses product-specific contracts, official catalog evidence, expert review logic, account credits, history, and source tracking so your AI agent can use a governed replacement engine instead of guessing from general web research.

Endpoint

Use the hosted Partglyph MCP endpoint:


https://app.partglyph.com/api-mcp/mcp

Configure your MCP client to use this endpoint with a Partglyph API key as a bearer token.

API Key

Create the key in the Partglyph app:


https://app.partglyph.com/settings/api-keys

Use the Partglyph MCP Access preset.

The preset is intended for normal MCP matching and includes:

Scope Purpose match:run Run account-governed product matches through Partglyph. history:read_own Read your own match history and result detail. credits:view_balance Check available account credits before a match.

Do not share the generated API key. API keys are shown only at creation time.

What You Can Ask

Use Partglyph MCP when you want your AI agent to work with industrial replacement questions such as:

  • “Find alternatives for UCP205 SKF.”

  • “Compare these replacement candidates and show the safest options first.”

  • “Only show drop-in results.”

  • “Include non-drop-in options too.”

  • “Open the detailed result for the best candidate.”

  • “Show my previous Partglyph matches.”

The AI client can read your chat, drawings, screenshots, documents, or links using its own available tools, then pass structured product information to Partglyph MCP for validation and matching.

What Partglyph Returns

Partglyph MCP returns structured, source-aware matching data that the customer AI client can present as tables, summaries, and detail views.

Normal result output can include:

  • requested input summary;

  • match status and result class;

  • part number and manufacturer candidates;

  • compatibility score;

  • product-specific flags and warnings;

  • technical comparison fields;

  • result details for deeper review;

  • history references for previous runs.

The AI client may format the response for the user, but it should preserve Partglyph statuses, warnings, result classes, and product-specific flags. It must not invent candidates, hide warnings, or strengthen a result beyond the evidence returned by Partglyph.

Result Controls

MCP match tools include non-drop-in candidates by default so users can see the engineering decision space.

Option Default Use include_non_drop_in true Set to false only when the user asks for drop-in results only. limit optional Set when the user asks for a smaller or larger result list.

Typical Workflow

A normal Partglyph MCP workflow is:

  • Identify the requested product family.

  • Prepare the required product fields.

  • Validate the prepared input.

  • Run the account-governed match.

  • Return a useful result table and summary.

  • Retrieve detailed result information when the user asks for it.

  • Retrieve match history when the user asks for previous runs.

History And Traceability

Successful MCP matches are treated as normal governed Partglyph runs.

This means:

  • account credits are handled through Partglyph user management;

  • the match appears in the user’s history;

  • result detail remains available through history tools;

  • the run is marked as coming from MCP for source tracking.

Accuracy And Review

Partglyph MCP is designed to reduce unchecked AI guessing in industrial replacement work.

It gives the customer AI client access to deterministic product contracts, official catalog evidence, expert matching logic, product-specific flags, and traceable history. Final engineering, procurement, and installation decisions should still be reviewed by qualified users before purchase or field use.