US Grants.gov Applicant Eligibility Change Watch

Changes to who may apply for a posted federal funding opportunity.

## Source
    Grants.gov public Search2 and FetchOpportunity APIs (public, logged-out)
    - https://api.grants.gov/v1/api/search2

• https://api.grants.gov/v1/api/fetchOpportunity

  Public, logged-out only — no API key, login, cookie or CAPTCHA.
  
    ## The eight delta types
    | value | fires when |
    |---|---|
    | `new-opportunity` | an opportunity appears in the watched set |
    | `forecast-to-posted` | status moves `forecasted` → `posted` |
    | `deadline-amendment` | the close date or synopsis response date moves |
    | `funding-range-change` | award ceiling, floor, estimated funding, number of awards or cost sharing changes |
    | `eligibility-change` | the set of eligible applicant types changes |
    | `opportunity-cancellation` | status moves into closed/archived/withdrawn, or the record is delisted |
    | `agency-program` | the agency, agency code or CFDA programme list changes |
    | `opportunity-document-change` | the attached document set changes |
  
    Leave `delta_types` empty for all eight. Every emitted delta carries `delta_kinds`
    naming which ones it satisfied.
  
    The last three need a per-opportunity detail read. Selecting any of them narrows
    the watched set to `max_detail_reads` opportunities, newest first, and uses a
    separate baseline — so a list-mode baseline is never diffed against a detail-mode
    read.
  
    ## Pricing
    | Event | Price (FREE tier) |
    |---|---|
    | Actor start (`apify-actor-start`) | $0.02 one-time |
    | Source check (`source-check`) | $0.10 |
  

| Per delta (grant-opportunity-delta) | $0.50 |

`source-check` ladder, GOLD-floored: **$0.10 / $0.09 / $0.08 / $0.067**.
    Charged once per run after the source enumeration terminates — seed runs and
    genuine-zero runs included. It is **not** charged when the source refused the read.

    The per-delta value event (`grant-opportunity-delta`) is the **primary** event and is priced at **$0.50 per delta** at the FREE tier. Ladder, GOLD-floored: **$0.50 / $0.45 / $0.40 / $0.335** (FREE / BRONZE / SILVER / GOLD-PLATINUM-DIAMOND). `source-check` still charges; it is no longer the headline event. Billed deltas are capped at 500 per run and any overflow is reported as a withholding rather than charged.

    ## Snapshot mode is a PREVIEW
    `watch_mode: false` returns at most **25 records**, hard-capped in code, labeled
    `output_mode: "preview"`. This actor sells change detection, not bulk export.

    ## Watch semantics
    The first watch run stores a private baseline and emits **zero** deltas. Billed
    deltas are capped at 500 per run; the overflow is a named PARTIAL withholding,
    never billed, and re-detected next run. When a read is truncated, `REMOVED` is
    suppressed rather than reported as a cancellation.

    ## Terminal states
    `NORMAL` · `SUCCEEDED-0: GENUINE_EMPTY` · `SUCCEEDED-0` (no change) · `PARTIAL` · `BLOCKED`.
    A zero with no evidence FAILS loud rather than exiting 0 quietly.

You name what this Actor watches

This Actor reads fields that only exist on the per-opportunity detail record, so it must issue one request per opportunity. It will not do that across the whole live set — instead it watches exactly what you name:

• opportunity_numbers — e.g. PAR-25-221. One bounded lookup each, exact-number match.
• agency_codes — e.g. DOI-NPS. Every live opportunity for that agency.

At least one is required. A run with neither fails immediately and says why, rather than quietly watching an arbitrary slice of the corpus and looking healthy.

Both bounds were proven to actually filter, control-vs-test on 2026-07-27 — oppNum takes 1,224 live opportunities to 1, agencies=DOI-NPS takes it to 16. A named opportunity that resolves to nothing is reported as NOT FOUND, never dropped in silence.

Why this matters for what you pay. Pool-wide, this field family changes on the order of a dozen times a year across all 1,224 live opportunities. Watching 40 opportunities you actually care about turns that into a signal; watching 150 arbitrary ones turns it into nothing.

One delta type per Actor

This Actor watches eligibility-change and nothing else. The other delta types in this family are separate Actors, each with its own price and its own baseline — so you pay for the feed you want and no other.

What "eligibility" means here, precisely

This Actor watches the structured applicant-type list — the coded set of who may apply, applicantTypes in the Grants.gov synopsis. It does not watch applicantEligibilityDesc, the free-text eligibility narrative.

That is deliberate, and it makes this Actor much quieter than the word "eligibility" suggests. Measured on 2026-07-27 across 375 revisions of 200 live posted opportunities: agencies modified the structured list 2 times and the free-text description 24 times — the wording moves about 12× more often than the rule it describes.

Agencies routinely reword the narrative — contact details, clarifications, formatting — without changing who is actually allowed to apply. Signing that text would bill you for copy-editing. So a delta from this Actor means the set of eligible applicant types changed, and if you need to know when the eligibility wording changed, this is not the Actor for it.

Run it WEEKLY, not daily

Schedule this Actor once a week. That is a recommendation about your bill, not about coverage — a weekly run reads the same opportunities and misses nothing.

This is a bounded, low-frequency feed by construction. Measured on 2026-07-27 across 200 of the 1,234 live posted opportunities, through this Actor's own classifier: 1.00 of this Actor's events per 100 watched opportunities per year. At the maximum bound this Actor accepts — 300 opportunities — that is about 3 deltas a year, or 0.06 a week.

A daily schedule pays seven source-check events a week to re-read a set that changes a few times a year: the check would be roughly 90% of your bill and buy no extra signal. On a weekly schedule the per-delta event carries about 19% of the bill and you see exactly the same changes. Override the cadence if you have a reason to.