US Government Contracts & Global Tenders Monitor

Track US federal contract awards from USAspending.gov and discover opportunities from official UK Contracts Finder, EU TED, and optional SAM.gov in one normalized dataset. USAspending, UK, and EU searches need no API key. Pay $0.002 per saved record and export buyers, award recipients, values, deadlines, qualification evidence, field-level changes, and official source URLs.

Use it to build a repeatable GovCon lead feed, find prime contractors after awards, qualify open tenders, compare procurement markets, or send a complete change-labelled portfolio into spreadsheets, CRMs, dashboards, and AI workflows.

Start Without an API Key

This small US-first run finds recent federal cybersecurity awards and saves at most 10 records:

{
  "sources": ["usaspending_awards"],
  "keywords": ["cybersecurity"],
  "noticeStatus": "active",
  "trackChanges": true,
  "monitorName": "us-cybersecurity-awards",
  "maxResults": 10
}

Supported Sources

Source	Coverage	API key required	Notes
USAspending.gov	US federal contract awards	No	Buyer agency, award recipient, amount, performance period, and modification date from the official API.
UK Contracts Finder	UK public-sector tenders and notices	No	Uses the official OCDS API.
EU TED	EU public procurement notices	No	Uses the official TED Search API.
SAM.gov	US federal opportunities	Yes	Requires your own SAM.gov public API key.

The default input includes USAspending awards, UK Contracts Finder, and EU TED. Add sam_gov only when you have a SAM.gov public API key.

What This Actor Extracts

• Source name and keyword used
• Stable source-scoped record key and source contract ID
• Contract title and description
• Buyer / contracting authority
• Award recipient / prime contractor when published
• Buyer country and region
• Notice type, stage, and procurement method
• Contract value and currency where available
• Published, last-modified, and deadline dates
• Award performance start and end dates
• Keyword match fields and a short match reason
• Deterministic fit score, fit reason, objective red flags, and recommended action
• Status such as active, closed, awarded, modified, or cancelled
• CPV, NAICS, PSC, or other classification codes
• Official contract or tender URL
• Scraped timestamp
• Named monitoring status, previous values, value/deadline deltas, and evidence-based change signals

The output is focused on official public procurement records and organization-level opportunity data. The Actor redacts email addresses and phone numbers from descriptions and does not intentionally collect personal contact lists, private account data, or non-public information.

Use Cases

• Monitor new government contracts and tenders by keyword.
• Build public-sector opportunity feeds for vendors and consultants.
• Track deadlines for proposal, bid, capture, and sales teams.
• Find award recipients for teaming, subcontracting, and competitor intelligence.
• Research procurement demand by country, buyer, category, value, and status.
• Feed dashboards, spreadsheets, CRMs, or internal lead-scoring workflows.
• Compare UK, EU, and US procurement markets in one normalized export.

Quick Start

US award and tender intelligence report

Use a decision profile when you want a review queue instead of only a raw tender export.

{
  "sources": ["uk_contracts_finder"],
  "keywords": ["software"],
  "noticeStatus": "active",
  "decisionProfile": {
    "preferredKeywords": ["software", "cloud", "cybersecurity"],
    "preferredRegions": ["London", "United Kingdom"],
    "preferredCategories": ["72000000", "IT services"],
    "excludedKeywords": ["hardware only"],
    "minimumContractValue": 100000,
    "minimumValueCurrency": "GBP",
    "minimumDaysToDeadline": 7
  },
  "maxResults": 10
}

The Actor saves the normalized records and writes a ranked TENDER_REPORT Markdown file. Tender rows receive bid-review actions; award rows can become partner_lead records when the published evidence fits. Scores do not predict win probability or replace human review.

Use this small input for a first run without any API key:

{
  "sources": ["uk_contracts_finder", "ted"],
  "keywords": ["software"],
  "noticeStatus": "active",
  "maxResults": 10
}

To search recent tenders from the default sources, you can also leave keywords empty:

{
  "sources": ["uk_contracts_finder", "ted"],
  "keywords": [],
  "noticeStatus": "active",
  "maxResults": 20
}

To include SAM.gov, add "sam_gov" to sources and provide samApiKey:

{
  "sources": ["sam_gov"],
  "keywords": ["cybersecurity"],
  "country": "CA",
  "samApiKey": "YOUR_SAM_GOV_API_KEY",
  "maxResults": 25
}

Input Fields

Field	Type	Default	Description
sources	array	USAspending + UK + EU	Official procurement sources to search. Only SAM.gov requires samApiKey.
keywords	string array	["software"]	Up to 10 optional search terms. Leave empty to return recent tenders from selected sources.
dateFrom	string	30 days ago	Publication start date in YYYY-MM-DD format. One run can cover at most 366 days.
dateTo	string	today	Publication end date in YYYY-MM-DD format.
country	string	empty	Optional country/region text filter. For SAM.gov, use a US state code such as CA or NY.
noticeStatus	active, all	active	Target active/open opportunities where supported, or include all available notices.
decisionProfile	object	optional	Preferred keywords, regions, categories, exclusions, value threshold/currency, and minimum deadline window used for deterministic triage.
maxResults	integer	10	Maximum total records saved across all selected sources and keywords.
samApiKey	string	empty	SAM.gov public API key, required only when sam_gov is selected.
trackChanges	boolean	true	Add a bounded private history and label every current row baseline, updated, unchanged, or partial.
monitorName	string	default	Independent watchlist name. Reuse the same name on scheduled runs.
observationHistoryLimit	integer	12	Prior observations retained per stable record, from 2-30.
proxyConfiguration	object	no proxy	Backward-compatible setting for saved tasks. Official APIs are called directly, so enabled proxy routing is rejected.

Output Overview

Each dataset item represents one normalized public procurement record.

Field group	Important fields
Source context	source, keyword, scrapedAt
Opportunity identity	recordKey, contractId, title, noticeType, stage, status, contractUrl
Buyer, recipient and location	buyerName, recipientName, buyerCountry, buyerRegion
Timing	publishedDate, lastModifiedDate, deadlineDate, performanceStartDate, performanceEndDate
Commercial detail	contractValue, currency, procurementMethod, classificationCodes
Description	description with email and phone-like text redacted where detected
Match evidence	matchedFields, matchReason
Decision support	fitScore, fitReason, redFlags, recommendedAction
Monitoring	monitoringStatus, changedFields, previous values, contractValueDelta, deadlineChangeDays, changeSignals, observation timestamps/count

Every run also writes:

• TENDER_REPORT: ranked Markdown review queue.
• RUN_SUMMARY: safe JSON outcome details, source warning counts, and spending-limit state. It never includes samApiKey.

Verified Output Example

The following record was saved from UK Contracts Finder by verification run jKAewnN2Nk1XmI0pQ on July 11, 2026. The description is shortened here for readability; the dataset retains the redacted source description.

{
  "source": "uk_contracts_finder",
  "keyword": "software",
  "recordKey": "uk_contracts_finder:ocds-b5fd17-5d26e35b-71a8-4b92-8178-8a47a5fef7be",
  "contractId": "ocds-b5fd17-5d26e35b-71a8-4b92-8178-8a47a5fef7be",
  "title": "CA18209 - The Mercian Trust Outsourced Payroll Service and HR and Payroll System",
  "buyerName": "The Mercian Trust",
  "buyerCountry": "England",
  "buyerRegion": "Walsall",
  "noticeType": "tender",
  "stage": "tender",
  "procurementMethod": "Open procedure",
  "contractValue": 400000,
  "currency": "GBP",
  "publishedDate": "2026-07-10T09:09:10.000Z",
  "lastModifiedDate": "2026-07-10T09:09:10.000Z",
  "deadlineDate": "2026-08-26T11:00:00.000Z",
  "status": "active",
  "classificationCodes": ["CPV:79631000 Personnel and payroll services"],
  "description": "The Trust is procuring a cloud-based HR and Payroll software system and outsourced payroll service.",
  "matchedFields": ["description"],
  "matchReason": "Keyword \"software\" matched description.",
  "fitScore": 80,
  "fitReason": "preferred keywords matched: software, cloud; preferred region matched: England; contract value published; 47 day(s) remain before deadline; active tender",
  "redFlags": ["No preferred category matched."],
  "recommendedAction": "review_now",
  "contractUrl": "https://www.contractsfinder.service.gov.uk/Notice/5fddd1b5-9a5f-48d0-9638-7f5377d5c06c-905563",
  "scrapedAt": "2026-07-11T07:43:26.670Z"
}

Pricing

Event	Price	When charged
Actor start	$0.00005 per GB	Platform start event; at least one event per run
contract-scraped	$0.002	Once per clean contract/tender record saved

Records are saved and charged atomically with contract-scraped. Free-user rows are still counted correctly, duplicate stable opportunity IDs are skipped before billing, and the Actor stops promptly when the user's spending limit is reached. The total result allowance is divided dynamically across selected source/keyword searches so the first source cannot consume the entire run by default; unused allowance flows to later searches.

Complete Change-Labelled Monitoring

Enable trackChanges and reuse one monitorName for scheduled runs. The Actor uses recordKey as the stable identity. Publication and modification dates are deliberately excluded from identity so an amendment remains the same opportunity or award.

Every run still returns the complete current export. Each row is labelled baseline, updated, unchanged, or partial, with changed fields, previous values, value and deadline deltas, fit-score movement, and signals such as recipient_changed, deadline_extended, or contract_value_increased. This differs from changed-only feeds that make a missing row ambiguous.

History is private to the Actor account, separated by a hashed monitor name, bounded to 2-30 observations per record, and committed only after the corresponding dataset row is saved. A missing value never silently overwrites the last known value. Absence from a capped search is not treated as proof that a tender closed or an award vanished.

If an official source returns multiple releases for the same stable opportunity during one run, the Actor keeps the newest lastModifiedDate version instead of billing duplicate amendments.

Date-only deadlines from sources such as TED are normalized to 23:59:59.999Z so an opportunity is not treated as expired at the start of its final calendar day. Exact source timestamps retain their original moment after conversion to UTC.

Reliability And Run Outcomes

• Official requests have bounded timeouts, polite per-host pacing, and at most three attempts.
• Only network errors, HTTP 408/425/429, and server errors are retried. Invalid input and authentication errors fail without wasteful retries.
• Contracts Finder pagination is restricted to its official HTTPS host. TED page repetition and malformed response shapes fail visibly.
• SAM.gov pagination is bounded, and API keys are redacted from logs, status messages, reports, and errors.
• A valid search with no matches succeeds with outcome: empty in RUN_SUMMARY.
• If one source fails and another completes, the run succeeds as partial with safe warnings.
• If every attempted official source fails, the run fails instead of pretending the outage was a genuine empty result.

Keyword evidence checks the longer description first, then title, classification, buyer, location, notice, stage, and procurement-method fields. matchedFields and matchReason explain why the record entered the result set.

Fit Scoring And Recommendations

The fit score is a transparent prioritization aid, not an AI prediction. It uses:

• preferred and excluded keyword matches in the published notice
• preferred buyer country/region and category/code matches
• published contract value against an optional same-currency threshold
• deadline availability and remaining time
• active/closed status and tender/award stage

review_now, review, monitor, and skip are deterministic recommendations. An expired/closed opportunity, an excluded keyword, or a published value below the configured minimum forces skip. Missing values and currency mismatches are flagged; the Actor never performs hidden exchange-rate conversion.

Every run writes TENDER_REPORT to the default key-value store and links it from the Output tab. Opportunities are ordered by fitScore, with red flags and official source links visible in the same table.

Tips For Better Results

• Start with maxResults: 10 to check the output before scaling.
• Keep one run within the 366-day publication-range limit; split larger historical backfills into separate tasks.
• Use focused keywords such as cybersecurity, software, facilities management, or medical equipment.
• Review matchedFields and matchReason when building qualification or alert workflows.
• Leave keywords empty when you want a broad recent-opportunity feed.
• Keep noticeStatus as active for open opportunities and switch to all for market research.
• Use SAM.gov only when you have a valid public API key from SAM.gov.

Known Limits

• Source APIs can differ in how they expose values, deadlines, locations, and classifications.
• Some notices do not publish contract values or exact deadlines.
• Fit scores depend only on fields exposed by the official source. They do not estimate competition, incumbent advantage, certification eligibility, procurement-law compliance, or win probability.
• Contract values are compared only when the record currency matches minimumValueCurrency; no currency conversion is performed.
• SAM.gov's public search response does not expose a separate modification timestamp, so lastModifiedDate falls back to publishedDate for that source.
• USAspending is post-award data, not an open-opportunity feed. Use optional SAM.gov for US bid opportunities and USAspending for buyer, recipient, amount, and award-change intelligence.
• Selecting SAM.gov without samApiKey is an input error, so a missing credential cannot look like a genuine empty result.
• SAM.gov may return a separate official description-API URL instead of description text. The Actor does not expose that URL as if it were a description.
• Country filtering checks buyer country/region for UK/TED and uses a US state-code API filter for SAM.gov.
• Source page ceilings protect runtime and official APIs. Very broad, low-signal searches may need narrower dates or keywords.

Responsible Use

Use this Actor only for lawful purposes and in compliance with the source websites' terms, robots.txt, applicable privacy laws, and local regulations. This Actor is designed for official public procurement records and does not intentionally collect personal contact lists, private account data, or non-public information.

Notes

• SAM.gov requires a user-provided public API key.
• EU TED Search API is public and does not require a key for published notices.
• UK Contracts Finder returns OCDS data under the UK Open Government Licence.