Claude · connected through Koddex MCP

Ask any question. Get an audited, traceable answer.

Through the Koddex MCP server, an AI agent reads your engineering data with the same scoped permissions and full audit trail as a human engineer.

claude.ai · workspace · koddex-mcp mcp connected

What's the operating temperature range of actuator J3 in WAND-EXO, and when was it last validated?

Ask Claude anything about your Koddex data…

Tools that aren't headless or API-first

Closed tools, no API

Legacy PLM/ERP suites have no headless API. You can't script them, can't integrate them, can't connect AI agents — every workflow ends in a copy-paste.

AI agents have no scoped access

An LLM either gets an admin password (terrifying) or no access at all (useless). There's no way to give an agent read-only, audited, attribute-level access to your engineering data.

External providers can't reach the data

Suppliers need a part spec, an auditor needs a verification trace — and you ship them a PDF or a leaky drive folder, with no way to revoke access or hide confidential fields.

The Koddex way — open, scoped, auditable

Headless, fully API-first

Every Koddex object is queryable through typed REST and GraphQL endpoints. Scripts, automations and dashboards plug in directly — no scraping, no exports.

MCP-native, scoped agent access

Koddex exposes its data model through Model Context Protocol. Agents authenticate, get typed scopes, and every read or write is logged with the same granularity as a human edit.

Granular external sharing

Invite a supplier or auditor with attribute-level scopes: share the dimensions, hide the cost; share the test results, hide the BOM. Revoke in one click. Every access logged.