Public procurement data · Europe

The AI database for
public tenders in Europe.

Our AI reads notices and tender documents and turns them into one record per contract: history, winner, price, criteria. Via API or directly in your AI.

2,000+
source portals
41
Countries
DEFRITPLESNLUKUA<cbc:ID>notice_4211.xml0,01 €+49 00000000siehe VergabeunterlagenCPV ????DUP #2<ted:Notice>N/APL-2026-118award=??p.32, PDFlinkeddeduplicatedcriteria extracted0,01 €820.000 €enrichedElektroinstallation,Stuttgartbuyer:Stadt Stuttgartstatus:COMPETITIONdeadline:12.09.2026value:820.000 €winner:—1 procedure · 6 notices
One data model. One API.

THE MARKET

Three reasons TED gives you no picture of the market.

01

Incomplete coverage

TED shows around 3,500 tenders per working day. Another 12,000 appear only on national portals, below the EU threshold or unreported to TED.

over 90 %

Share of contracts that never reach TED in some markets

02

Fragmented notices

One tender appears on TED as several notices with no link between them: call, change, award. The same buyer appears under a dozen spellings.

49 %

Share of TED notices that are actually calls for competition

03

Poor data quality

Contract values as €0.01 or missing. Phone numbers like +49 00000000. Award criteria in running text instead of fields. Hard-to-read titles, typos in the buyer's name.

41

Countries, 41 formats, no shared data model

TED fulfils a publication duty, not your need for information. TenderHub does.

THE SOLUTION

One record per contract. Complete, linked, clean.

ONLY AT TENDERHUB

The tender documents, read and structured

Specifications, forms and annexes are read and delivered as fields in the record.

Criteria

Deadlines

Quantities

Certificates

Conditions

One procedure, one record

National and EU notice of the same tender in one record, every change dated.

Award and winner

Winner, contract value and bidder count per award. Estimated and actual value kept separate.

Competitor profiles

Losing bidders are captured too: who wins what, where and against whom.

Buyer register

München, Muenchen, MÜNCHEN: one record per buyer with all procedures.

Expected tenders

Expiring frameworks and prior information notices as a preview of upcoming procedures.

Project groups for construction

All trades of a construction project as one group instead of single hits.

Below-threshold tenders

Tenders below the EU threshold included, switchable on demand.

Region and radius

Every address geocoded, radius search around your location.

Directly in your AI. TenderHub.eu is available as an MCP server. Connect the data to the AI assistant you already use and ask in plain language.

ChatGPT

Claude

Copilot

Gemini

Le Chat

PORTAL COVERAGE

DEEUATFRNLPLITESBECHSEDKNOFIPTGRCZROHRIEUKUA+ 19 more

2,000+ source portals across Europe, aggregated through the central national platforms of 41 countries, all available in the beta.

ASK THE DATA

Ask our AI anything about European tenders.

Any question, answered with procurement data from 41 European countries, now in beta. Ask in plain language, get numbers back. Try it on your own market.

TenderHub AI

41 COUNTRIES · BETA

What do French authorities pay for facility cleaning?

Across France, the median winning price is €23.40 per m² and year, ranging from €18.10 to €29.80. Prices rose 4.2% over the last year.
1,240 awards · 2024–2026 · national portals

Ask about your buyers, prices, competitors…

Free. GDPR compliant. No newsletter.

FOR DEVELOPERS

Five notices. One record.

A primary school is being extended and drywall work is needed. On TED and in the German notice service that is five XML documents over eight months. Below you see what TenderHub.eu makes of them, as a merge request: the source on the left, our record on the right.

Prior information112 Mar 2026
Contract notice230 Jul 2026
National copy330 Jul 2026
Corrigendum425 Aug 2026
Award notice520 Nov 2026
MergedMerge raw data from five notices into one procurement record
OverviewCommits 9Changes 5 ·
TenderHub.eu

All countries, all sources, one deduplicated JSON data model over REST.

Example procedure with a fictional buyer. The data errors are real: every class is measured in the corpus and handled by the API exactly as shown here.

COMPARISON

How TenderHub compares.

TenderHub.eu
TED-only providers
Open aggregators
Coverage
41 countries. National sources plus TED, below-threshold included
TED only, roughly a third of the market
Varies by country, gaps common
Information depth
Notice plus AI-extracted criteria as structured fields
The notice only
The notice only
Data model
One unified model across all countries
TED schema as published
Raw per-source formats
Freshness
Daily, under 24 hours from publication
Daily
Often weeks behind, irregular
Data quality
Deduplicated, cleaned, AI enriched
As published
Duplicates and gaps
Navigability
Procedure level: all notices of a tender linked into one story
Notice level
Notice level, unlinked

PROVEN IN PRODUCTION

The team behind TenderHub.

We are Ciconia Systems from Stuttgart, a team of software and AI engineers. With Vergabepilot.AI we built the tender search for German companies. TenderHub.eu turns it into a data foundation for all of Europe.

01

The challenge

Germany's procurement market is scattered across hundreds of portals. No bid team can watch them all by hand.

02

What we built

Vergabepilot.AI, the first AI-based semantic search for German public tenders. Search, matching and AI summaries on top of one clean data engine.

03

The result

Over 14,000 users, 300 German portals monitored daily, AI summaries and matching in production since 2024.

TenderHub takes the same data engine European: 41 countries, one API.
Vergabepilot.AI – Die KI für Ausschreibungen

Users on Vergabepilot.AI

14,000+

German portals monitored

300+

Daily ingestion since launch

2024

Marko Jeftic

Marko Jeftic

CO-FOUNDER & CEO

Has been building digital and AI products for over 10 years. Responsible for product and customers at TenderHub.eu.

Daniel Holderbaum

Daniel Holderbaum

CO-FOUNDER & CTO

Built the data engine behind Vergabepilot.AI. Responsible for architecture and operations at TenderHub.eu.

Prof. Dr. Christian Bartelt

Prof. Dr. Christian Bartelt

SCIENTIFIC ADVISOR

Professor of machine learning at TU Clausthal. Brings the latest research into our AI.

BETA

Waitlist for beta access

Over the coming months, TenderHub.eu opens step by step to everyone on the waitlist, with the first access going to users who want to start early. Describe as concretely as possible how you would use the data or the API: this helps us prepare your access properly. We will contact you personally as soon as it is ready.

Join the waitlist
STEP 1 OF 2

FAQ

Frequently asked questions