Finding government contracts: tenders from TED, the UK, Latin America and India in one feed
How bid teams and B2B sellers pull open tenders and contract awards from nine official procurement portals plus India's CPPP and GeM, with real inputs, a daily alert script in Python and exact prices.
The actor referenced in this article. Pay only for results delivered.
Government buyers publish every contract they put out to bid, but they publish it in dozens of places. A firm that sells IT services in Germany, the UK and India has to watch TED, Contracts Finder, Find a Tender, the CPPP and GeM, each with its own search, its own alert email and its own export format, if it has one at all. Bid teams end up checking portals by hand, missing notices that closed in ten days, and pasting what they find into a spreadsheet that nobody else can filter.
Try it live: Public Tenders Scraper. Pay per result delivered.
Two actors turn that into one table. The Public Tenders Scraper reads nine official procurement portals across Europe, the UK, Australia, Latin America, Ukraine and Canada. The India Tenders Scraper reads India’s Central Public Procurement Portal, the 28 state and UT portals it collects, and GeM bids. Both return one row per notice and can run every morning on a schedule.
Which actor covers which market
| Market | Actor | Sources | Tenders | Awards with winner |
|---|---|---|---|---|
| EU and EEA | Public Tenders Scraper | TED (27 EU countries plus Norway, Iceland, Switzerland and others) | Yes | Yes, where published |
| United Kingdom | Public Tenders Scraper | Contracts Finder, Find a Tender | Yes | Yes |
| Australia | Public Tenders Scraper | AusTender | No, contract notices only | Yes |
| Brazil, Colombia, Chile | Public Tenders Scraper | PNCP, SECOP II, ChileCompra | Yes | Yes |
| Ukraine | Public Tenders Scraper | Prozorro | Yes | No winner in the search data |
| Canada | Public Tenders Scraper | CanadaBuys | Open tenders | No |
| India | India Tenders Scraper | CPPP central, 28 state and UT portals, GeM | Active tenders only | No |
For US federal contracts, see our posts on SAM.gov and Companies House data and the USASpending API.
International tenders: nine portals in one run
Each source is searched once per keyword, and every row has the same columns whichever portal it came from: buyer_name, title, cpv_codes or classification, estimated_value, currency, deadline_date, notice_url, and for awards winner_name, award_value and award_date.
A bid pipeline for four EU countries looks like this:
{
"countries": ["DE", "FR", "PL", "ES"],
"keywords": ["road construction", "IT services"],
"noticeType": "tender",
"daysBack": 30,
"maxItems": 100
}
EU and EEA country codes narrow TED to buyers in those countries. GB, AU, BR, CO, CL, UA and CA switch on that country’s own source. noticeType takes tender, award or any. daysBack counts back from today, or set an exact dateFrom and dateTo. maxItems is per source and keyword, so two sources and three keywords at 100 can return up to 600 rows. A notice that two keywords find is delivered once.
TED, Contracts Finder, PNCP, SECOP and Prozorro search their own full text. Find a Tender, AusTender, ChileCompra and CanadaBuys are read day by day from their open feeds and filtered by your keyword, where every word must appear in the title, description, buyer or category. Search national portals in the buyer’s language for best recall.
Contract awards: who won and for how much
Awards are where competitor research and partner hunting start. This input lists who won cleaning contracts in the UK and Australia in the last month:
{
"sources": ["uk_contracts_finder", "uk_find_a_tender", "au_austender"],
"keywords": ["cleaning services"],
"noticeType": "award",
"daysBack": 30,
"maxItems": 500
}
A trimmed award row from TED, as it appears in the README:
{
"source": "eu_ted",
"country": "PL",
"notice_id": "783456-2026",
"notice_type": "award",
"cpv_codes": ["90910000", "90500000", "90511000", "90512000"],
"published_date": "2026-11-11",
"award_date": "2026-11-09",
"award_value": 61668,
"currency": "PLN",
"winner_name": "DOBRY KLIMAT Andrzej Sergiel",
"winner_country": "PL",
"notice_url": "https://ted.europa.eu/en/notice/-/detail/783456-2026"
}
The winner is filled when the source publishes it. In the README test runs, 209 of 209 UK Contracts Finder awards, 90 of 90 AusTender contracts, 205 of 211 Find a Tender awards and 127 of 142 TED award notices had it. Group the rows by winner_name and you have a list of companies that already sell to government in your category: competitors to watch, or subcontracting partners to call.
India: CPPP, state portals and GeM
The India Tenders Scraper reads the whole active list of each source, then applies your filters before a row is delivered, so you pay only for tenders that match. This input pulls big civil works in one state that close in the next two weeks:
{
"sources": ["state"],
"states": ["Uttar Pradesh"],
"categories": ["Civil Works"],
"minValueInr": 10000000,
"closingWithinDays": 14,
"sortBy": "closing_soon",
"maxItems": 200
}
CPPP rows carry the numbers a bidder checks first as plain numbers: tender_value_inr, emd_inr and tender_fee_inr, plus bid_submission_end_at, days_to_close, inviting_officer, inviting_address and portal_url. GeM rows carry the bid number, the items asked for, quantity, bid_type and document_url. Here is a trimmed central tender from the README:
{
"source": "central",
"tender_id": "2026_MES_795641_1",
"title": "ANNUAL MAINT CONTRACT OF SOLAR PLANT AT 09 LOCATIONS UNDER GE (S) JAIPUR",
"ministry_or_state": "Ministry of Defence",
"tender_category": "Works",
"tender_value_inr": 1493000,
"emd_inr": 29860,
"bid_submission_end_at": "2026-11-14T18:00:00+05:30",
"inviting_officer": "CWE JAIPUR",
"days_to_close": 2.7
}
A state, value or category filter on central tenders reads each candidate’s full record, which is slower, so add keywords to narrow the candidates first.
A daily tender alert in Python
This script runs both actors for yesterday’s notices and writes one CSV sorted by deadline:
import csv
from datetime import date
from apify_client import ApifyClient
client = ApifyClient("YOUR_APIFY_TOKEN")
def notices(actor, run_input):
run = client.actor(actor).call(run_input=run_input)
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
if "_type" not in item: # skip summary, info and error rows
yield item
alerts = []
for n in notices("themineworks/public-tenders-scraper", {
"sources": ["eu_ted", "uk_contracts_finder", "uk_find_a_tender"],
"keywords": ["software", "IT services"],
"noticeType": "tender",
"daysBack": 1,
"maxItems": 200,
}):
alerts.append({
"source": n["source"],
"title": n.get("title_english") or n.get("title", ""),
"buyer": n.get("buyer_name", ""),
"deadline": n.get("deadline_date", ""),
"value": f'{n.get("estimated_value", "")} {n.get("currency", "")}'.strip(),
"link": n.get("notice_url", ""),
})
for t in notices("themineworks/india-tenders-scraper", {
"sources": ["central", "state", "gem"],
"keywords": ["software"],
"publishedWithinDays": 1,
"maxItems": 200,
}):
alerts.append({
"source": f'india_{t["source"]}',
"title": t["title"],
"buyer": t.get("organisation") or t.get("ministry_or_state", ""),
"deadline": t["bid_submission_end_at"][:10],
"value": f'{t["tender_value_inr"]} INR' if "tender_value_inr" in t else "",
"link": t.get("portal_url") or t.get("document_url", ""),
})
alerts.sort(key=lambda a: a["deadline"] or "9999")
if alerts:
with open(f"tenders_{date.today().isoformat()}.csv", "w", newline="", encoding="utf-8") as f:
writer = csv.DictWriter(f, fieldnames=list(alerts[0].keys()))
writer.writeheader()
writer.writerows(alerts)
print(f"{len(alerts)} new tenders")
Run it once a day, or save both inputs as tasks in Apify Console with a morning schedule and send the datasets to Google Sheets or Slack. To be sure you see each notice only once, compare runs on source plus notice_id, and on cppp_id or gem_bid_id for India.
A second lead list: trade fair exhibitors
Award winners tell you who sells to government in your category. If you also want the companies in that industry that do not bid yet, a trade fair’s exhibitor list is another honest source. The Trade Show Exhibitors Scraper takes an exhibitor list link from Map Your Show, Swapcard, Messe Frankfurt, RX and five other directory platforms and returns one row per exhibitor with booth, country, website and, where the fair publishes them, email and phone. It costs $4.50 per 1,000 exhibitors on the Free plan and $3.60 on Gold, plus $0.003 per run.
What it costs
Both actors bill per row delivered plus a start fee, and the start fee is charged only once the first result is found, so a day with no matching tenders is free.
| Actor | Free | Bronze | Silver | Gold and above | Start fee per run |
|---|---|---|---|---|---|
| Public Tenders Scraper, per 1,000 notices | $3.60 | $3.15 | $2.70 | $2.52 | $0.001 |
| India Tenders Scraper, per 1,000 tenders | $1.80 | $1.55 | $1.30 | $1.08 | $0.003 |
A notice found by two keywords, notices past maxItems, India tenders that miss your filters, and the summary, info and error rows are never charged. Full India tender records cost nothing extra. Put all your keywords and sources in one run so the start fee is paid once.
Limits worth knowing
- No documents. Neither actor downloads tender documents; rows link to where they are.
- Winners are not everywhere. Some TED notices withhold the winner or value for confidentiality, Prozorro’s search has no winner, and CanadaBuys gives open tenders only.
- AusTender is awards only. It publishes contract notices, not open tenders.
- Caps per search. Up to 10,000 notices per source and keyword. Prozorro shows at most 1,000 matches, and ChileCompra, which has no keyword search in its API, opens up to 2,000 tenders per keyword.
- India is active tenders only. Award results (AOC) sit behind a captcha and are not covered. About one in five buyers publishes no estimated value, and a value filter drops those tenders.
- Speed varies. International searches take seconds, ChileCompra minutes. In India, full records are read at about 100 a minute.
Related
Explore the scraper referenced in this article: inputs, outputs, and pricing, then run it on Apify.
Frequently asked questions
Do these actors download the tender documents? +
No. The Public Tenders Scraper links each notice's official page in notice_url and does not download documents. India Tenders Scraper rows link the tender on its source portal (portal_url), and GeM rows link the bid document (document_url). You open the documents there.
Can I see who won a contract? +
For the international sources, yes: set noticeType to award and rows carry winner_name, award_value and award_date where the source publishes them. Prozorro's search never includes the winner. For India, award results are not covered, because the portals put them behind a captcha.
What does a daily alert cost? +
A daily run that delivers 100 new international notices costs $0.361 on the Free plan ($0.36 plus the $0.001 start fee) and $0.253 on Gold. 100 Indian tenders cost $0.183 on Free and $0.111 on Gold. A day with nothing new costs nothing, because the start fee waits for the first result.
Which language should my keywords be in? +
The buyer's language for national portals, for example medicamentos for Brazil and Chile. TED searches every EU language and adds an English headline in title_english. On India's CPPP a keyword matches the title, description, reference, category or organisation, and on GeM it runs GeM's own bid search.
Company verification from official registries: Europe, US states, Brazil and India
How to check a company's status, registration, directors and address against official registers in eight European countries, seven US states, Brazil and India, with real inputs, a Python script and exact prices.
Bundesanzeiger Scraper: German Company Filings Without a Data Subscription
Pull German company filings from the Bundesanzeiger by name: annual accounts, insolvency proceedings, ad hoc disclosures and voting rights notices. No login, no API key.
Federal Contracts, Campaign Finance & UK Company Records as Structured JSON (No Login)
Three official government data sources, SAM.gov contract opportunities, FEC campaign finance, and UK Companies House, turned into clean JSON APIs for GovCon capture, compliance, and political research. No browser, no login, pay only for results delivered.
CourtListener API: How to Search US Court Records and Case Law Programmatically
CourtListener exposes 10M+ court opinions and dockets via a free REST API. Here is how to query it, what the rate limits actually are, and when a scraper is faster.