Partsgraph MCP¶

The agent-ready parts data layer. Partsgraph turns product catalogs into a canonical Parts Graph and serves it to AI agents over MCP. The public demo catalogue carries provenance-backed part records with datasheet-sourced specs, compatibility checks, and BOM building for reference builds. The company model is enterprise: the same graph technology serves your own product catalog over MCP, agent-readable pages, and shopping feeds.

Server type: Remote Streamable HTTP (hosted, keyless)
Endpoint: https://partsgraph.ai/api/mcp
Auth: None
Tools: 11 verified by live probe (server v1.4.0)
Registry: ai.partsgraph/catalog
Scope: demo catalogue, 35 provenance-backed part records, 3 reference builds

Why This Matters for Operators¶

Parts sourcing drowns in PDFs and vendor silos: a spec change means re-checking compatibility, and a BOM means pricing dozens of line items. Partsgraph makes the loop agent-native. search_parts finds real parts with provenance-backed attributes, get_provenance shows where each spec came from, check_compatibility validates part pairs before procurement, and build_bom composes a priced bill of materials. The demo catalogue is the public sandbox; the enterprise product applies the same tools to your own catalog.

Tools & Capabilities¶

Tool	What it does
`search_parts`	Search real electronic and mechanical parts with provenance-backed attributes
`get_part`	Full record for one part with datasheet-sourced specs
`get_provenance`	Provenance trail for a part's attributes
`compare_parts`	Side-by-side part comparison
`list_categories`	Browse the parts taxonomy
`list_builds`	Reference builds in the demo catalogue
`check_compatibility`	Compatibility checks between parts
`build_bom`	Compose a priced bill of materials
`search`, `fetch`, `view_bom`	Generic catalog search, retrieval, and BOM views

Installation¶

Connect directly; no account or API key required for the demo catalogue:

{
  "mcpServers": {
    "partsgraph": { "type": "http", "url": "https://partsgraph.ai/api/mcp" }
  }
}

Configuration¶

None for the public demo. The endpoint is Streamable HTTP with public initialize and tools/list. Enterprise deployments serve your own catalog as a Parts Graph over the same protocol.

Business Relevance¶

Sourcing: search and compare real parts with datasheet provenance instead of vendor PDFs.
Compatibility risk: check part pairs before committing to a BOM.
Costing: priced BOMs for early build-cost estimates.
Catalog strategy: the enterprise product makes your own catalog agent-readable.

Integration with CorpusIQ¶

Partsgraph's structured part and BOM output pairs with CorpusIQ connectors for procurement workflows: push BOM line items into Airtable or Notion for tracking, log supplier quotes in your CRM, and render BOM cost dashboards from the returned priced data.

Limitations¶

The public endpoint serves a demo catalogue (35 part records, 3 reference builds), not the full parts universe.
Enterprise deployments for your own catalog are a commercial engagement with partsgraph.ai.
No public repository; the platform is hosted.