EU TED Public Procurement & Tenders Scraper — Self-Healing Apify Actor & MCP Server (eu-ted-procurement-tenders-mcp)

{
  "id": "ted:15980-2026",
  "publication_number": "15980-2026",
  "publication_date": "2026-01-12",
  "title": "Ireland - Security software development services - Provision of Security Operations Centre and Cyber Defence services to the Rehab Group",
  "buyer_name": "The Rehab Group",
  "country_code": "IRL",
  "cpv_codes": ["72212730", "72212732", "72611000"],
  "notice_type": "cn-standard",
  "total_value": 500000.0,
  "ted_url": "https://ted.europa.eu/en/notice/-/detail/15980-2026",
  "pdf_url": "https://ted.europa.eu/en/notice/15980-2026/pdf",
  "ppe_event": "ted_notice_extracted",
  "billable": true,
  "network_mode": "live_eu_ted_v3_api",
  "extracted_at": "2026-10-28T19:15:00Z"
}

Keywords: EU TED Tenders, European Public Procurement, CPV Code Tender Monitor, B2B Government Contracts, MCP Server, eu-ted-api-v3, eforms-ubl-xml

---

Why B2B Bid Teams, GovTech Vendors & RevOps Use This Actor (ROI vs. DIY)

Public sector procurement across the 27 EU Member States and EEA exceeds EUR 2 trillion annually on Tenders Electronic Daily (TED). Following the European Commission's migration to the TED Search API v3 (api.ted.europa.eu/v3/notices/search) and eForms UBL XML (Regulation 2019/1780), legacy HTML scrapers and v2 integrations fail or return unnormalized multilingual dictionaries. This Actor converts complex CPV, ISO country, and keyword filters into clean, English-prioritized B2B contract records ready for CRM ingestion or AI agent tool-calling via the Apify MCP Server.

Approach	Setup & Maintenance Effort	eForms & Multilingual Normalization	MCP / AI Agent Ready	Cost Structure
In-House Script (DIY)	18-30 hrs (TED v3 Expert Query syntax + 24-language map + UBL XML parser)	Breaks when notices omit English titles or switch eForms subtypes	Manual wrapper required	EUR 900+ engineering setup
Enterprise Tender Portals	Multi-week sales onboarding & annual seat lock-in	High	Closed proprietary web UI	EUR 350-EUR 1,200/mo minimum
This Actor (eu-ted-procurement-tenders-mcp)	0 hrs (Instant Apify API, CLI, or MCP)	3-Tier Self-Healing (TED API v3 + eForms UBL XML + HTML/JSON-LD)	Native JSON & MCP Schema	Pay-Per-Event ($0.002 / notice)

---

30-Second Copy-Paste Quickstart

Run a live EU TED procurement query filtering by keyword, 8-digit CPV codes (72000000 IT Services, 48000000 Software Packages), and 2-letter or 3-letter ISO buyer countries:

python3 workspaces/actors/eu-ted-procurement-tenders-mcp/main.py \
  --query "cybersecurity" \
  --cpv "72000000,48000000" \
  --countries "DE,FR,NL,IE" \
  --published-after "2026-01-01" \
  --max-items 5

Or pass an inline JSON configuration payload matching INPUT_SCHEMA.json:

python3 workspaces/actors/eu-ted-procurement-tenders-mcp/main.py \
  --input-json '{"searchQuery":"cloud","cpvCodes":["72000000","48000000"],"countries":["DEU","FRA","NLD","IRL","CYP"],"publishedAfter":"2026-01-01","maxItems":10}'

Run the built-in deterministic self-test and unit test suite:

python3 workspaces/actors/eu-ted-procurement-tenders-mcp/main.py --self-test
python3 -m unittest discover -s workspaces/actors/eu-ted-procurement-tenders-mcp -p "test_*.py" -v

---

3-Tier Self-Healing EU TED Extraction Architecture

1. Pre-Flight Query Builder & ISO/CPV Normalizer (build_ted_expert_query):
   • Automatically maps 2-letter ISO-3166-1 alpha-2 country codes (DE, FR, NL, IE, CY, EL/GR, UK/GB, SE, PL, etc.) to the mandatory 3-letter ISO-3166-1 alpha-3 codes required by TED v3 (DEU, FRA, NLD, IRL, CYP, GRC, GBR, SWE, POL).
   • Sanitizes 8-digit CPV codes (normalize_cpv_codes), stripping optional check digits (48000000-8 -> 48000000).
   • Validates publishedAfter dates (YYYY-MM-DD or YYYYMMDD) and compiles valid TED v3 Expert Search queries (publication-date >= 20260101 AND buyer-country IN (DEU FRA) AND classification-cpv IN (72000000) AND FT ~ "cybersecurity").
2. Tier 1 Primary — Official EU TED Search API v3 (POST https://api.ted.europa.eu/v3/notices/search):
   • Queries the official European Union Publications Office REST v3 endpoint with pagination, exponential backoff on HTTP 429/5xx, and deterministic multilingual resolution (extract_multilingual_text) across all 24 official EU languages (prioritizing English eng, then deu, fra, nld, esp, ita).
3. Tier 2 Fallback — Official EU TED eForms UBL XML Parser (parse_ted_eforms_xml):
   • If the TED Search API gateway experiences transient maintenance, the Actor automatically falls back to parsing official eForms UBL XML notices (https://ted.europa.eu/en/notice/{publication_number}/xml), extracting <efbc:NoticePublicationID>, <cac:PartyName>, <cac:ProcurementProject>, <cbc:EstimatedOverallContractAmount>, and <cbc:ItemClassificationCode listName="cpv">.
4. Tier 3 Self-Healing HTML & JSON-LD Fallback (extract_with_fallbacks):
   • Multi-tier regex and application/ld+json selector chain with strict BrokenSelectorError enforcement so corrupt or empty responses are never billed.

---

Input & Output Schema Reference

Input Parameters (INPUT_SCHEMA.json)

• searchQuery (string, optional): Keyword (e.g. "software", "cloud", "cybersecurity") or raw TED v3 Expert Search expression.
• cpvCodes (array of strings, optional): 8-digit CPV classification codes (e.g. ["72000000", "48000000"]).
• countries (array of strings, optional): Buyer countries in 2-letter (DE, FR, CY) or 3-letter (DEU, FRA, CYP) ISO format.
• publishedAfter (string, optional): Publication date floor in YYYY-MM-DD or YYYYMMDD format (default "2026-01-01").
• maxItems (integer, required, default 25, min 1, max 500): Maximum number of procurement notices to extract.

Output Record Fields

• publication_number: Official TED publication identifier (e.g. "15980-2026").
• publication_date: Notice publication date (YYYY-MM-DD).
• title: Normalized tender title (English preferred with automatic EU multilingual fallback).
• buyer_name: Official legal name of the contracting authority.
• country_code: 3-letter ISO-3166-1 alpha-3 buyer country code (DEU, FRA, IRL, etc.).
• cpv_codes: Deduplicated array of 8-digit CPV classification codes.
• notice_type: Official TED notice type (cn-standard, can-standard, pin-only, etc.).
• total_value: Disclosed estimated or total contract value (numeric or null when not disclosed by the buyer).
• ted_url: Canonical URL to the notice on ted.europa.eu.
• pdf_url: Direct link to the official notice PDF.
• ppe_event: "ted_notice_extracted".
• billable: true when the record passes schema validation.
• network_mode: "live_eu_ted_v3_api" (or "live_eu_ted_eforms_xml" on Tier 2 failover).
• extracted_at: ISO-8601 UTC extraction timestamp.

---

Scope & Compatibility (What's Included / What's Not Included)

• What's Included:
  • Zero-dependency Python 3.11+ Apify Actor (main.py) querying the official EU TED Search API v3 (POST https://api.ted.europa.eu/v3/notices/search) with automatic 2-to-3 letter ISO country code normalization, 8-digit CPV code sanitization, multilingual field extraction, and eForms UBL XML fallback.
  • Direct links to the official TED HTML notice (ted_url), multilingual PDF (pdf_url), and machine-readable eForms XML (xml_url).
  • Full Apify Store and Model Context Protocol (MCP) compatibility (INPUT_SCHEMA.json, .actor/actor.json, .actor/dataset_schema.json, .actor/output_schema.json, Dockerfile, and test_actor.py).
• What's Not Included (Anti-Refund Warranty Boundary):
  • Does not bypass password-protected national e-Tendering portals (such as external buyer login portals linked inside tender specifications for submitting sealed bids).
  • Does not populate total_value when the contracting authority has legally omitted or redacted the estimated contract value in the official TED notice (total_value will be null in those cases).
  • Does not cover municipal micro-tenders below EU publication thresholds that are not published in the Supplement to the Official Journal of the European Union (OJ S / TED).