Portfolio · Reference implementation
A public reference for AI-Ready European data.
NACEBridge is a free, open utility that maps the EU NACE Rev 2.1 backbone down to the 5th-digit national activity codes used across the Union. It is also — deliberately — a portfolio piece: a working example of how modern public-sector data should ship in the age of assistants.
Built by
AI Readiness & AI Projects.
We help organisations become legible to AI: MCP servers, OpenAPI specs, Schema.org graphs, Markdown mirrors, retrieval-friendly information architecture, and evaluation harnesses. NACEBridge is our public reference of what that looks like end-to-end when applied to a real European dataset.
tech.teksoma.comPowered by
European market data pipeline.
MI ingests, normalises and versions authoritative datasets from Eurostat and each EU27 national statistical institute. NACEBridge pulls its full taxonomy — EU classes, national subclasses, and 24-language translations — from MI on a monthly revision cycle.
mi.teksoma.comWhat AI Readiness means, concretely
Every fact on this site is available to a human in a browser, to a developer through a REST call, and to an AI assistant through a native protocol — from the same canonical store, with the same citations, versioned together.
Streamable HTTP · JSON-RPC 2.0 · Claude, Cursor, ChatGPT, Devin.
REST spec for /api/v1/* — importable into any Custom GPT.
Discovery index following the llmstxt.org convention.
Flattened corpus for embedding pipelines.
Full TOC + citations + download URLs in one JSON call.
Machine-readable citation graph.
NACE Rev. 1 (1970) → Rev. 2.1 (2025).
Per-chapter plain Markdown mirrors — /handbook/chapters/{slug}/md.
Organization + WebSite + Dataset on every page; Book on the handbook.
GPTBot, ClaudeBot, PerplexityBot, Google-Extended, CCBot — all explicitly allowed.
Because migrants shouldn't need a translator
A Portuguese founder should be able to find her CAE code in France in seconds, not weeks.
Because data should teach itself
If an AI can't cite you, you don't exist. Machine-readable provenance is table stakes.
Because Europe deserves better plumbing
Twenty-seven official languages, one taxonomy — served over open standards, free of charge.
If your organisation is thinking about MCP, OpenAPI, RAG readiness, or shipping AI projects that actually reach production, TΣKSOMΛ Technologies takes on a small number of engagements per quarter.