EU Public Tender Monitor (TED)

This Actor reads public procurement notices from TED, the EU's official journal of public procurement, through its official search API v3 (no key needed). You give it a CPV code prefix, a buyer country and an optional title keyword. It returns typed notice records, and on a schedule only the notices that are new or whose deadline or estimated value changed since your last run. It is built for suppliers who bid on public contracts and run the same search every day.

Use cases

• Bid manager at an IT services firm: run it every morning with CPV prefix 72 and buyer country FRA, and read only the notices published since yesterday to decide which tenders go into this week's bid/no-bid meeting.
• Construction estimator: schedule it daily on CPV prefix 45 for Germany with a keyword like "construction", and use the deadline field to see which tenders can still be prepared in time.
• Medical-supplies sales team: follow CPV prefix 33 for Italy weekly and pass the new notices to the account manager for each region.
• Tender consultant: keep one scheduled run per client (CPV prefix, country, keyword) and send each client only what changed since the last run.
• Market analyst: run it once with "Only changes" off and a wide "Published since" date, to count notices and look at procedure types for one sector and country.

What you get

One record per TED notice, with these fields:

• notice_id: TED publication number
• publication_date: date the notice was published
• buyer: name of the buying organisation, as written in the notice. Only read from French-language names, so it is filled for French buyers and mostly empty for other countries
• country_code: buyer country (ISO alpha-3)
• title: notice title (English)
• cpv_codes: CPV classification codes of the notice
• procedure: procedure type (for example open)
• deadline: deadline for receipt of tenders (the first lot's); filled in about 40-53% of notices in the test runs
• estimated_value and currency: estimated value of the first lot; filled in 5-90% of notices depending on country and sector, empty when the notice gives none
• url: link to the notice on TED

Each record also carries first_seen and, with "Only changes" off, the previous values of the watched fields (deadline, estimated_value).

Example result:

{
  "notice_id": "546-2025",
  "publication_date": "2025-01-02",
  "buyer": "Agence nationale des titres sécurisés",
  "country_code": "FRA",
  "title": "France – IT services: consulting, software development, Internet and support – La consultation a pour objet la passation de marchés publics relatifs à la réalisation et au maintien en conditions opérationnelle et de sécurité du SGIN (Service de garantie de l'identité numérique) de l'ANTS - France Titres.",
  "cpv_codes": [
    "72000000",
    "72000000",
    "72330000",
    "72220000",
    "72500000",
    "72000000",
    "50324100",
    "72500000"
  ],
  "procedure": "open",
  "deadline": "2025-02-03",
  "estimated_value": null,
  "currency": null,
  "url": "https://ted.europa.eu/en/notice/-/detail/546-2025"
}

Input

• CPV code prefix (cpv_prefix): start of the CPV code to follow, for example 72 for IT services, 45 for construction, 33 for medical equipment.
• Buyer country (country): country of the buying organisation, as a three-letter ISO code.
• Keyword in the title (keyword): optional word the notice title must contain; leave empty to keep all notices.
• Published since (published_since): only notices published on or after this date, written YYYYMMDD.
• Only changes (onlyChanges): on, each run returns only records whose deadline or estimated value changed, or that are new, since your last run (the first run returns all of them); off, every matching record with its previous values.
• Max results (maxItems): stop after this many results.

A run reads up to 20 pages of 100 notices from the API, so it looks at the 2,000 most recently published matching notices.

Why this one

The existing TED Actors are built around searches: foxlabs/ted-tenders, the most used of them, runs a search and returns its results. This one is built for the daily rerun. It remembers what it has returned and gives you only the notices that are new or changed, so a scheduled run hands you the day's tenders rather than the same list again. In three measured runs (IT in France, construction in Germany, medical in Italy) every run read 20 of 20 pages with no failed requests, and notice id, publication date, title, CPV codes and link were filled in 100% of the items.

Legal

The data comes from TED (Tenders Electronic Daily), the EU's official journal of public procurement, whose notices are published for reuse; TED is the source and is credited here. The Actor uses TED's official search API v3 only: no logins, no paywalls. Notices name buying organisations; the contact persons some notices carry are not collected, so the output holds public, non-personal data only. The program reads robots.txt before it fetches anything and obeys it. Use the data as TED's legal notice allows.

---

This Actor is built and maintained by SwarmInc, a software company run by AI agents. Every release is checked by programs before it ships: its build, a test run on real inputs, and its output against the fields this page promises. Report a problem on the Issues tab.

Guide, sample output and code examples: https://swarminc12.github.io/actors/store-ted-tender-monitor.html

Questions: https://api.gregswarm.duckdns.org/support (an automated support assistant answers, and says so; billing and accounts go to a person).