Canada Government IT Tender Monitor

Find open Canadian federal IT procurement opportunities from CanadaBuys and export them as clean, structured data.

This Actor helps SaaS, cloud, cybersecurity, data, AI, and IT services vendors monitor public-sector tender opportunities without manually searching CanadaBuys every day.

What this Actor does

• Downloads official CanadaBuys tender notice CSV data.
• Filters for IT-related opportunities such as software, cloud, cybersecurity, data, AI, TBIPS/SBIPS, database services, and digital services.
• Scores each matching notice with a deterministic fitScore from 0 to 100.
• Outputs normalized records with buyer, dates, category, source identifiers, matched keywords, summary, source URL, and contact details.
• Works with Apify datasets, API exports, schedules, webhooks, and downstream AI-agent workflows.

Best use cases

• Monitor Canada federal IT RFPs and tender opportunities.
• Build a daily watchlist for government sales teams.
• Feed procurement opportunities into CRM, Slack, email, or internal bid pipelines.
• Track cybersecurity, cloud, AI, data, and software-related public-sector opportunities.
• Create a structured procurement intelligence feed for agents and automations.

Data source

The Actor uses official CanadaBuys open data files published through the Government of Canada open data ecosystem.

Default source:

https://canadabuys.canada.ca/opendata/pub/openTenderNotice-ouvertAvisAppelOffres.csv

Important scope note: this first version focuses on CanadaBuys federal tender notice data. It does not claim full coverage of all provincial, municipal, or external procurement portals that can appear in the CanadaBuys web search experience.

Input

Recommended default input:

{
  "source": "open",
  "status": "open",
  "minScore": 50,
  "maxItems": 100,
  "maxRowsScanned": 0,
  "dropExcluded": true,
  "includePastClosingDates": false
}

Main input fields

Field	Type	Description
source	string	open, new, or complete. Use open for current open opportunities.
status	string	open, expired, cancelled, or all.
keywordsInclude	array	Optional custom IT keywords. Leave empty for the default IT keyword set.
keywordsExclude	array	Optional custom exclusion keywords. Leave empty for defaults.
dateFrom	string	Optional publication date lower bound, YYYY-MM-DD.
dateTo	string	Optional publication date upper bound, YYYY-MM-DD.
provinces	array	Optional buyer province filters, for example Ontario.
minScore	integer	Minimum IT-fit score from 0 to 100. Default: 50. Use 40 for broader recall.
maxItems	integer	Maximum dataset items to output. Default: 100.
dropExcluded	boolean	Drop notices matching exclusion keywords. Default: true.
includePastClosingDates	boolean	Include notices whose closing date is already in the past. Default: false.

Output

Each dataset item is one normalized tender notice.

Key output fields:

Field	Description
id	Stable hash based on source notice, solicitation, and amendment identifiers.
sourceNoticeId	CanadaBuys/MERX reference number when available.
solicitationNumber	Solicitation number.
amendmentNumber	Amendment number.
title	Tender title.
buyerName	Contracting entity.
endUserName	End user entity when available.
status	Tender status.
publishedDate	Publication date.
closingDate	Tender closing deadline.
procurementCategory	CanadaBuys procurement category codes such as SRV, GD, or SRVTGD.
procurementMethod	Procurement method.
gsinDescription	GSIN description when available.
unspscDescription	UNSPSC description when available.
summary	Short normalized description snippet.
fitLabel	strong_it_fit, possible_it_fit, or weak_it_fit.
fitScore	Deterministic IT relevance score from 0 to 100.
matchedKeywords	Keywords that matched the tender.
excludedKeywordsMatched	Exclusion terms that matched, if retained.
sourceUrl	Original notice URL, or a CanadaBuys tender-notice URL constructed from the source reference number when the CSV URL field is empty.
attachments	Attachment URLs or names when available.
contact	Contact name, email, and phone when available.
fetchedAt	UTC timestamp for this Actor run.

Example output item

{
  "fitScore": 72,
  "fitLabel": "strong_it_fit",
  "title": "APN - CYBERSECURITY ASSESSMENT SOURCE LIST",
  "buyerName": "Defence Construction Canada / Construction de Defense Canada",
  "closingDate": "2026-07-08T14:00:00",
  "matchedKeywords": ["cyber", "cybersecurity", "information technology"],
  "sourceUrl": "https://canadabuys.canada.ca/en/tender-opportunities/tender-notice/cb-951-14703007"
}

Data source, freshness, and affiliation

Each run downloads the latest available CanadaBuys open tender CSV for the selected source mode.

Current scope:

• CanadaBuys tender notices.
• Primarily Canadian federal open tender data available in the official open-data files.
• Not provincial, municipal, MERX, SEAO, or BC Bid coverage yet.

Affiliation and verification:

• This Actor is not affiliated with CanadaBuys or the Government of Canada.
• Always verify tender details, attachments, deadlines, amendments, and bid instructions on the official source URL before bidding.
• Source update timing depends on the official CanadaBuys data publication process.

Source modes

open

Best default. Processes currently open CanadaBuys federal tender notices.

new

Useful for frequent polling and alert-style workflows. Processes newly published tender notices.

complete

Useful for historical analysis. Processes the complete CanadaBuys tender notice file from 2022-08-08 onward. This is heavier than open.

Tips

• Use minScore: 50 for the default cleaner feed.
• Use minScore: 40 for broader recall.
• Use minScore: 70 for a stricter high-confidence IT feed.
• By default, past closing date-times are skipped even when the source status is still Open.
• Use source: new with Apify schedules and webhooks for alerting workflows.
• Add custom keywordsInclude for your niche, for example zero trust, SOC, SIEM, Oracle, Salesforce, ServiceNow, or Microsoft 365.
• Add custom keywordsExclude when a broad keyword creates false positives.

Limitations

• This Actor is procurement intelligence, not legal or bid advice.
• Source completeness depends on the official CanadaBuys open data files.
• The baseline score is deterministic keyword/classification scoring, not a legal assessment of whether your company should bid.
• Long tender descriptions are truncated only in summary; the full normalized description is still available in description.
• Attachment parsing and AI summarization are not included in v1. They are planned as future enrichment features.

Planned improvements

• Saved presets for cybersecurity, cloud/software, TBIPS/SBIPS, and AI/data.
• Better amendment grouping and duplicate handling.
• New-tender alert mode examples.
• Optional AI enrichment for tender summaries and company-fit scoring.
• Attachment parsing.
• Future expansion to SAM.gov and TED/EU using the same normalized schema.

Quality controls

• scoreExplanation explains why a tender matched, so users can audit fitScore instead of trusting a black box.
• strictMode: true narrows the feed to opportunities with stronger IT evidence in the title or classification fields.
• focusArea lets users quickly narrow the feed to cybersecurity, cloud/SaaS, data/AI, software licensing, IT professional services, or TBIPS/SBIPS opportunities.
• dedupeMode: latestBySolicitationNumber keeps the latest record per solicitation number by default. Use none if you need every source notice/amendment row.

Company-fit matching

Use companyFitMode and companyProfileKeywords when you want the Actor to score or filter tenders for a specific vendor profile.

Example:

{
  "companyFitMode": "filter",
  "companyProfileKeywords": ["SOC", "penetration testing", "cybersecurity assessment"],
  "focusArea": "cybersecurity",
  "strictMode": true
}

• companyFitMode: score keeps all normal results and adds company-fit fields.
• companyFitMode: filter keeps only results matching at least one company keyword.
• sortBy: companyFitScore puts best company-profile matches first.

Large source safety

For source: complete, the Actor streams the official CSV and applies a default maxRowsScanned safety limit of 20000 rows unless you provide a different value. This avoids downloading and scanning the complete historical file blindly during normal runs. Use date filters, stricter filters, and explicit row limits when exploring historical data.