GrantConnect Grant Awards Scraper & Monitor — Apify Store Overview

Actor: stefano_seggio/australia-grantconnect-monitor · Actor ID: gt7wS4T0uFRXDz49n · Store URL: https://apify.com/stefano_seggio/australia-grantconnect-monitor Version: 2.0 (.actor/actor.json) — latest published patch per CHANGELOG.md is 2.0.1 (2026-09-07)

---

Executive Summary & Business Use Case

This Actor extracts the Grant Awards register on GrantConnect (grants.gov.au) — the official, whole-of-government record of every grant a Commonwealth (Australian federal) entity has awarded since December 2017. GrantConnect is operated by the Department of Finance; under the Commonwealth Grants Rules and Principles every non-corporate Commonwealth entity is legally required to publish each grant it awards, and each subsequent variation, within 21 days of the agreement taking effect. The result — recipient name, ABN, awarding agency, dollar value, purpose, funding program, delivery location and grant term — is public data, but GrantConnect itself exposes it only through a manual search form and a report capped at 50,000 rows, with no API, no CSV feed and no email alert for awards (its native notifications cover open funding opportunities only, not awards already made). This Actor turns that register into structured JSON/CSV, with server-side filtering by keyword, category, recipient, ABN, value range, Grant Opportunity ID and date window, plus a delta ("Only new") mode that returns only what changed since the previous run.

Three concrete, source-grounded use cases: grant consultants and bid writers query agency, grantProgram, goId, valueAudNumber and selectionProcess to see which agency just funded organisations like their client under which program, so they can approach fresh winners about variation work or pitch the next funding round. Government relations, lobbying and integrity researchers filter on recipientAbn, agency, event_type and isOneOffAdHoc to get alerted the moment a named competitor, client interest or portfolio agency receives a new or varied award, and to flag large non-competitive ("ad hoc") awards for scrutiny — precisely the pattern the Australian National Audit Office has flagged, given ANAO's own count of 215,000+ grants worth A$118 billion (plus A$36 billion in variations) on the register. Suppliers to grant recipients (construction, professional services, equipment vendors) use deliveryState, deliveryPostcode, valueBand and purpose to identify who just received money to build, buy or deliver something in their region, and reach out within days of the award instead of months later when it surfaces in an annual report.

Every use case above maps to a field that actually exists in this Actor's dataset schema — there is no per-transaction pricing signal beyond the award's own face value (valueAudNumber), no procurement-bid data, and no data on Grant Opportunities (open funding rounds still accepting applications); this Actor covers awards already made, not rounds still open.

Technical Features & V2 Architecture Highlights

Delta engine with named, cross-run persistence. When onlyNew is true, the Actor writes each delivered award's lastUpdatedIso watermark to a private, named key-value store, australia-grantconnect-monitor-state-<deltaStateName> (the state name defaults to a fingerprint of the run's filter set, so unrelated schedules never collide unless you deliberately share a deltaStateName). Per the input schema's own description of onlyNew: it "remembers every award it has delivered (per filter set, in a named key-value store) and returns only awards that are new, varied or updated since. The walk stops as soon as it reaches already-known records, so a scheduled run costs a few pages instead of the whole register." State is written only for records that were actually stored in the dataset, and records are delivered oldest-first within a run, which is what makes a mid-run spending-limit cutoff or platform migration crash-safe: the next run simply resumes rather than silently losing an award. Version 2.0.1 fixed two edge cases in this engine specifically: a backlog floor, so that when a run stops at maxItems with more matching rows still below it, the next run walks all the way down through the already-known block to that floor before trusting an early stop (previously it could stall short of undelivered rows); and a baseline floor, so a cheap, capped first run doesn't cause every later run to slow-drain the entire historical archive.

Real, domain-specific event types — deliberately no CLOSED. The dataset's event_type field is constrained to exactly three values: NEW_LISTING (a brand-new award), AWARD_VARIATION (a separate amendment record GrantConnect publishes when an award's value or term changes, e.g. GA270901-V1 varying base award GA270901), and UPDATED (an award already delivered by a previous run whose lastUpdated stamp moved — a correction or republish, delta mode only). There is no CLOSED event on this actor, unlike register actors that model listings which literally close: a GrantConnect award is a permanent record of money already paid, not an open listing that can be withdrawn, so the domain-correct signal for "this record changed" is UPDATED, and for "this is a new document tied to an old award" it's AWARD_VARIATION — not a status transition. sortBy defaults to Last Updated rather than Publish Date specifically to make variations visible: GrantConnect stamps a variation record with the original award's publish date, so ordered by publish date a fresh variation can sit 100,000 rows deep on the day it appears, while ordered by Last Updated it's on page one.

Envelope, not a 18-field UMS claim. This Actor's own dataset_schema.json defines a 6-field integration envelope — record_id, event_type, scraped_at, is_new, source_url, data_source — which the README states is "identical across all of this developer's public-register Actors, so one webhook parser serves them all." That envelope sits alongside roughly 84 award-specific fields (identity, listing, detail, recipient, delivery location, agency contact, provenance), for 90 properties total in the schema — the README's own "80+ fields" claim undercounts it slightly. No 18-field baseline is documented anywhere in this Actor's schema files, so that figure is not asserted here.

No browser, no proxy, no login. Per actor.json's meta.templateId (ts-crawlee-cheerio), this is a plain-HTTP Cheerio crawler, not a headless browser — it runs on 256–1024 MB of memory (minMemoryMbytes / maxMemoryMbytes), needs no residential proxy, and authenticates nothing because GrantConnect's award register is fully public. fetchDetail (default true) controls whether the Actor opens each award's own page for the ~30 detail-only fields (ABN, delivery location, purpose, program, Grant Opportunity link, selection process, confidentiality flags, agency contact) or stops at the cheaper listing-only summary; maxConcurrency (default 5, max 10) controls how many detail pages are fetched in parallel.

Input Schema & JSON Configuration Example

All fields below are read directly from .actor/input_schema.json.

Field	Type	Default	Description
keyword	string	"" (prefill)	Full-text search over grant title, purpose and activity (also accepts a GA ID such as GA578886). Matching mode below.
keywordMatch	string (enum: AllWord, AnyWord, ExactPhrase)	AllWord	How the keyword(s) must match.
categories	array of string (29 category-code enum)	[]	GrantConnect's 29 top-level categories. Select several to get their union. The record's own 'category' field shows the finer sub-category label (e.g. 'Medical Research').
agencyNameContains	string	—	Case-insensitive substring match on the awarding agency, e.g. 'Health' or 'Infrastructure'. Applied client-side (GrantConnect's public form has no agency parameter), so combine it with a category, keyword or date window to keep the walk short.
recipientName	string	—	Substring match on the recipient's legal name, e.g. 'Monash University' or 'Regional Express'.
recipientAbn	string (pattern: 11 digits, spaces optional)	—	Exact match on the recipient's 11-digit Australian Business Number (spaces optional). Returns every Commonwealth grant that entity has received.
minValueAud	integer	—	Only awards worth at least this amount. Tip: 1000000 narrows the 360,000-record register to roughly 28,000.
maxValueAud	integer	—	Only awards worth at most this amount.
goId	string (pattern: GO + digits)	—	All awards made under one Grant Opportunity (funding round), e.g. GO6105.
dateType	string (enum: Publish Date, Approval Date, Start Date, End Date, Current, Closed)	Publish Date	Which date the 'from'/'to' window applies to. 'Current' = grants running today and 'Closed' = grants whose term has ended (both ignore the window).
dateFrom	string (absolute or relative datepicker)	—	Absolute (2026-07-01) or relative ('7 days', '3 months', '1 year' back from today, Canberra time). Leave empty for the whole archive (December 2017 onwards).
dateTo	string (absolute or relative datepicker)	—	Absolute or relative upper bound of the window. Leave empty for 'today'.
oneOffAdHoc	string (enum: any, yes, no)	any	Ad hoc grants (no competitive round) are the ones probity and integrity teams watch most closely.
aggregateGrantAward	string (enum: any, yes, no)	any	Aggregate awards bundle several recipients into one record ('Multiple').
eventTypes	array of string (enum: NEW_LISTING, AWARD_VARIATION, UPDATED)	["NEW_LISTING", "AWARD_VARIATION", "UPDATED"]	NEW_LISTING = a brand-new award; AWARD_VARIATION = a variation/amendment record (value or term changed, e.g. GA270901-V1); UPDATED = an award you already received whose 'Last Updated' stamp moved (delta mode only).
onlyNew	boolean	false (prefill true in the Console UI)	Remembers every award it has delivered (per filter set, in a named key-value store) and returns only awards that are new, varied or updated since. The walk stops as soon as it reaches already-known records, so a scheduled run costs a few pages instead of the whole register.
sortBy	string (enum: Last Updated, Publish Date, Relevance)	Last Updated	'Last updated' (recommended, and required for reliable monitoring): variations and corrections of old awards keep their original publish date, so only this order surfaces them. 'Relevance' needs a keyword.
deltaStateName	string (pattern: alphanumeric/hyphen, 1–30 chars)	— (fingerprint of filters)	Optional label for the memory of a monitoring task. Defaults to a fingerprint of your filters, so different schedules never interfere. Set the same name on two tasks to make them share one memory.
resetState	boolean	false	Forget every previously delivered award for this delta state and re-baseline (the run then returns up to 'Max records' most recently updated awards).
maxItems	integer (1–100000)	100	Hard cap on delivered records (and therefore on cost: you pay per record). In delta mode anything beyond the cap is simply delivered by the next run.
fetchDetail	boolean	true	Opens each award's page (one extra request) for recipient ABN and location, delivery location, purpose, program, Grant Opportunity link, selection process, confidentiality flags and the agency contact. Off = listing-only summary records (cheaper 'result-summary' price).
maxConcurrency	integer (1–10)	5	How many award pages are fetched at once. GrantConnect tolerates 10 comfortably; 5 keeps a 500-record run under two minutes.
dateRange	string (hidden, pattern: 24h, 7d, 30d)	—	Kept for tasks created with version 1: '24h', '7d' or '30d' are interpreted as 'Date from'.

Valid JSON configuration example

{
  "keyword": "renewable energy",
  "keywordMatch": "AnyWord",
  "categories": ["211", "301"],
  "minValueAud": 250000,
  "dateType": "Publish Date",
  "dateFrom": "90 days",
  "eventTypes": ["NEW_LISTING", "AWARD_VARIATION"],
  "onlyNew": true,
  "sortBy": "Last Updated",
  "deltaStateName": "renewables-watch",
  "maxItems": 500,
  "fetchDetail": true,
  "maxConcurrency": 5
}

This example combines category 211 (Environment, Energy and Resources) and 301 (Science and Technology), a $250,000 floor, a rolling 90-day publish window, and delta mode under a named state (renewables-watch) so a schedule reusing this exact input only ever receives new activity.

Output Dataset Sample & Data Dictionary

Fields below are taken directly from .actor/dataset_schema.json. Every property is nullable/optional except the six required fields (record_id, event_type, scraped_at, is_new, source_url, gaId).

Integration envelope (identical across this developer's public-register Actors):

Field	Type	Description
record_id	string	Same value as gaId — stable across runs (e.g. GA578886, GA270901-V1 for a variation).
event_type	string (enum: NEW_LISTING, AWARD_VARIATION, UPDATED)	NEW_LISTING, AWARD_VARIATION or UPDATED (a previously delivered award whose Last Updated stamp moved).
scraped_at	string	ISO-8601 UTC timestamp of this extraction.
is_new	boolean	true when this award was never delivered by a previous run of this delta state.
source_url	string	The official GrantConnect record page.
data_source	string	Attribution string (CC BY 3.0 AU).

Identity:

Field	Type	Description
gaId	string	GrantConnect's own award identifier (example: GA578886).
gaUuid	string | null	Stable identifier from the record URL.
isVariation	boolean	Is variation.
variationNumber	integer | null	1 for GA270901-V1.
baseGaId	string	The original award this record belongs to (equals gaId for a base award).
variesGaId	string | null	Varies GA ID.
variesUrl	string | null	Varies URL.
title	string | null	Grant Activity title as shown on the listing.

Listing:

Field	Type	Description
agency	string | null	Awarding agency name as recorded at publication.
category	string | null	GrantConnect sub-category label, e.g. 'Medical Research'.
publishDate	string | null	Publish date, raw site string (example: 4-Sep-2026).
grantTerm	string | null	Grant term, raw site string (example: 4-Sep-2026 to 30-Nov-2027).
valueAud	string | null	Value AUD, raw site string (example: $225,000.00).
recipientName	string | null	Recipient name.
lastUpdated	string | null	Last updated, raw site string (example: 4-Sep-2026 4:16 pm (ACT Local Time)).
publishDateIso	string | null	YYYY-MM-DD.
lastUpdatedIso	string | null	UTC ISO-8601 instant converted from Canberra local time — the register's only time-of-day field and the delta engine's change key.
valueAudNumber	number | null	Numeric value for sorting, summing and thresholds.
valueBand	string | null	<10k, 10k-100k, 100k-1M, 1M-10M, >=10M.
grantStartDate	string | null	Grant start, raw site string.
grantEndDate	string | null	Grant end, raw site string.
grantStartDateIso	string | null	Grant start (ISO date).
grantEndDateIso	string | null	Grant end (ISO date).
grantTermDays	integer | null	Grant term, in days.
daysUntilGrantEnd	integer | null	Negative once the term has ended — a renewal/re-tender signal.
isCurrent	boolean | null	true while today falls inside the grant term.
financialYear	string | null	Financial year (published), example: 2026-27.
grantStartFinancialYear	string | null	Financial year (start).
recipientEntityType	string | null	company, local_government, university, aboriginal_corporation, incorporated_association, trust, government, multiple or other — derived from the legal name.

Detail (populated when fetchDetail: true):

Field	Type	Description
detailFetched	boolean	false when fetchDetail was off or the award page could not be retrieved (see detailError).
detailError	string | null	Detail error.
agencyCurrentName	string | null	Agency name on the award page — reflects machinery-of-government renames.
approvalDate	string | null	Approval date, raw site string.
approvalDateIso	string | null	Approval date (ISO).
variationPublishDate	string | null	Variation publish date, raw site string.
variationPublishDateIso	string | null	Variation publish date (ISO).
variationDate	string | null	Variation date, raw site string.
variationDateIso	string | null	Variation date (ISO).
oneOffAdHoc	string | null	One-off / ad hoc, raw site string.
isOneOffAdHoc	boolean | null	Is one-off / ad hoc.
aggregateGrantAward	string | null	Aggregate grant award, raw site string.
isAggregate	boolean | null	Is aggregate.
aggregateReason	string | null	Aggregate reason.
numberOfAwardsAggregated	integer | null	Number of awards aggregated.
gstInclusive	boolean | null	true when the award page marks the value 'GST inclusive where applicable'.
pbsProgramName	string | null	PBS program name.
grantProgram	string | null	Grant program.
grantActivity	string | null	Grant activity.
purpose	string | null	Purpose.
goId	string | null	The originating Grant Opportunity (funding round), when linked.
goUuid	string | null	GO UUID.
goTitle	string | null	GO title.
goUrl	string | null	GO URL.
internalReferenceId	string | null	Internal reference ID.
selectionProcess	string | null	Selection process (example: Demand Driven).
confidentialityContract	string | null	Confidentiality — contract, raw site string.
isContractConfidential	boolean | null	Contract confidential.
confidentialityReasonContract	string | null	Confidentiality reason — contract.
confidentialityOutputs	string | null	Confidentiality — outputs, raw site string.
isOutputsConfidential	boolean | null	Outputs confidential.
confidentialityReasonOutputs	string | null	Confidentiality reason — outputs.

Recipient:

Field	Type	Description
recipientAbn	string | null	Recipient ABN, raw site string (example: 46 101 325 642).
recipientAbnNormalized	string | null	11 digits, no spaces — the join key for ABR, ASIC, CRM and credit data.
recipientAbnValid	boolean | null	ABN checksum valid.
abrLookupUrl	string | null	ABR lookup URL.
recipientSuburb	string | null	Recipient suburb.
recipientTownCity	string | null	Recipient town/city.
recipientPostcode	string | null	Recipient postcode.
recipientState	string | null	Recipient state.
recipientCountry	string | null	Recipient country.

Delivery location:

Field	Type	Description
deliverySuburb	string | null	Delivery suburb.
deliveryTownCity	string | null	Delivery town/city.
deliveryPostcode	string | null	Where the grant is delivered — can differ from the recipient's address (e.g. 'Multiple').
deliveryState	string | null	Delivery state (example: ACT, VIC).
deliveryCountry	string | null	Delivery country.

Agency contact:

Field	Type	Description
agencyContactName	string | null	Agency contact name.
agencyContactPhone	string | null	Agency contact phone.
agencyContactEmail	string | null	Agency contact email.

Sample dataset record

{
  "record_id": "GA578886",
  "event_type": "NEW_LISTING",
  "scraped_at": "2026-09-06T10:15:57.202Z",
  "is_new": true,
  "source_url": "https://www.grants.gov.au/Ga/Show/937de059-5cbb-415f-bc83-bcb5619e1379",
  "data_source": "GrantConnect (grants.gov.au), Australian Government Department of Finance, CC BY 3.0 AU",
  "gaId": "GA578886",
  "gaUuid": "937de059-5cbb-415f-bc83-bcb5619e1379",
  "isVariation": false,
  "baseGaId": "GA578886",
  "title": "The Activity is to support low volume and/or new commercial airline routes to regional and remote communities...",
  "agency": "Department of Infrastructure, Transport, Regional Development, Communications, Sport and the Arts",
  "category": "Transport",
  "publishDate": "4-Sep-2026",
  "publishDateIso": "2026-09-04",
  "lastUpdated": "4-Sep-2026 4:16 pm (ACT Local Time)",
  "lastUpdatedIso": "2026-09-04T06:16:00.000Z",
  "valueAud": "$225,000.00",
  "valueAudNumber": 225000,
  "valueBand": "100k-1M",
  "grantTerm": "4-Sep-2026 to 30-Nov-2027",
  "grantStartDateIso": "2026-09-04",
  "grantEndDateIso": "2027-11-30",
  "grantTermDays": 453,
  "daysUntilGrantEnd": 450,
  "isCurrent": true,
  "financialYear": "2026-27",
  "recipientName": "Regional Express Pty Ltd",
  "recipientEntityType": "company",
  "recipientAbn": "46 101 325 642",
  "recipientAbnNormalized": "46101325642",
  "recipientAbnValid": true,
  "abrLookupUrl": "https://abr.business.gov.au/ABN/View?abn=46101325642",
  "recipientSuburb": "MASCOT",
  "recipientState": "NSW",
  "recipientPostcode": "2020",
  "deliveryState": "NSW",
  "deliveryPostcode": "2020",
  "approvalDateIso": "2026-08-21",
  "selectionProcess": "Demand Driven",
  "isOneOffAdHoc": false,
  "isAggregate": false,
  "isContractConfidential": false,
  "pbsProgramName": "ITRDCSA 26/27 2.3 Air Transport",
  "grantProgram": "Airservices Australia Enroute Charges Payment Scheme",
  "purpose": "To provide regional and remote communities with access to essential air services...",
  "goId": "GO6105",
  "goTitle": "Airservices Australia Enroute Charges Payment Scheme",
  "goUrl": "https://www.grants.gov.au/Go/Show?GoUuid=706c1e7b-90df-4e26-b6e3-757c464a717f",
  "agencyContactEmail": "EGMO@infrastructure.gov.au",
  "detailFetched": true
}

Multi-language Integration Snippets

All three snippets call the same endpoint with the same real input: category 231 (Health, Wellbeing and Medical Research) at minValueAud 500,000, in delta mode, capped at 500 records.

cURL

curl -X POST "https://api.apify.com/v2/acts/stefano_seggio~australia-grantconnect-monitor/run-sync-get-dataset-items?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "categories": ["231"],
    "minValueAud": 500000,
    "onlyNew": true,
    "maxItems": 500
  }'

Python (apify-client)

from apify_client import ApifyClient

client = ApifyClient(token="$APIFY_TOKEN")

run = client.actor("stefano_seggio/australia-grantconnect-monitor").call(
    run_input={
        "categories": ["231"],
        "minValueAud": 500000,
        "onlyNew": True,
        "maxItems": 500,
    }
)

for award in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(award["gaId"], award["event_type"], award["recipientName"], award["valueAudNumber"])

Node.js (apify-client)

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: '$APIFY_TOKEN' });

const run = await client.actor('stefano_seggio/australia-grantconnect-monitor').call({
    categories: ['231'],
    minValueAud: 500000,
    onlyNew: true,
    maxItems: 500,
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();
for (const award of items) {
    console.log(award.gaId, award.event_type, award.recipientName, award.valueAudNumber);
}

Pricing Model Explanation

This Actor is pay-per-event, verified live via the Apify API:

Event	Price	When it fires on this Actor
result	$0.003 per record	A record delivered with the full detail page fetched — i.e. fetchDetail: true and the award's detail page was successfully retrieved (detailFetched: true), carrying all ~30 detail-only fields (ABN, delivery location, purpose, program, GO link, selection process, confidentiality flags, agency contact) on top of the listing fields.
result-summary	$0.001 per record	A listing-only record — either because fetchDetail was set to false for the run, or because fetchDetail was true but that specific award's detail page could not be fetched (detailFetched: false, with detailError set, e.g. a NOT_FOUND on a withdrawn page). In both cases the record still carries the full listing fields, just none of the detail-page fields.
apify-actor-start	$0.00005	Once per run, regardless of how many records it delivers.

The two-tier result / result-summary split on this Actor is specifically about whether the award's own detail page was fetched, not about "new vs. previously seen" — that distinction is handled entirely by delta mode, separately from pricing tier. Worked examples from the Actor's own README: a filtered 500-record pull with fetchDetail: true costs about $1.50; the same 500 records listing-only (fetchDetail: false) would cost about $0.50; a daily monitor that finds 40 new or varied awards costs roughly $0.12/day; a quiet monitoring run that finds nothing new costs the $0.00005 start fee only.

Delta mode and billing. When onlyNew: true, an award that GrantConnect has not changed since the last run for that deltaStateName is never delivered at all — the crawler's own walk stops as soon as it reaches records it already has recorded in the named key-value store. An unchanged record is therefore not created as a dataset item and is not a billable event of any kind; it is not billed at $0, it simply never exists as an item for that run. Only awards that are genuinely new (NEW_LISTING), amended (AWARD_VARIATION) or whose lastUpdated stamp moved (UPDATED) are written to the dataset and billed at the result or result-summary rate above, plus the one-time per-run start fee.

---

Source files read: .actor/actor.json, .actor/input_schema.json, .actor/dataset_schema.json, README.md, CHANGELOG.md — all from C:\Users\Stef\apify-portfolio\australia-grantconnect-monitor.