PubMed literature

Watch PubMed literature via NCBI E-utilities and generate citation reports

Job: PubMed citation metadata, literature-update reports, and new-PMID alerts through official NCBI E-utilities - not procurement bids and not Google News RSS.

Outputs are descriptive citation metadata. No abstracts / no full text. Not medical advice. Not affiliated with NCBI or PubMed.

From $4.00 / 1,000 pubmed metadata rows ($0.004 per metadata row). Schema defaults report/export ON - set a cap so you do not surprise-charge.

Cheap monitor signal vs report+export audit

Path Mode Report/export Cap hint What you get
Cheap signal runMode monitor OFF $0.30 Up to 10 × $0.03 new-publication alerts
Literature update runMode audit ON $7 $2 report + $5 export when non-empty

Monitor baseline / unchanged poll = $0. No matches = no report/export charge.

Cheap signal sample:

{
  "searchTerms": ["machine learning healthcare"],
  "runMode": "monitor",
  "monitorKey": "starter-machine-learning-healthcare",
  "emitRawRows": false,
  "generateReport": false,
  "generateExport": false,
  "maxChargeUsd": 0.30
}

Audit sample (prefill theme):

{
  "searchTerms": ["cancer immunotherapy"],
  "fromDate": "2026",
  "runMode": "audit",
  "emitRawRows": false,
  "generateReport": true,
  "generateExport": true,
  "maxChargeUsd": 7
}

Open PubMed Literature Watch → for a cheap signal use monitor + report/export OFF; for a literature update keep audit + report/export ON under a $7 cap. Citation metadata only — not full text.

Official PubMed E-utilities, not procurement bids and not news RSS

Use this page when the job is PubMed citation metadata, literature-update reports, systematic-review citation exports, or saved-query new-publication alerts. Use TED, SAM.gov & Grants Bid Alerts Scraper when the job is TED, SAM.gov, and Grants.gov bid queues. Use Google News Scraper when the job is query-based Google News RSS headlines and article URLs.

This Actor TED, SAM.gov & Grants Bid Alerts Google News Scraper
Intent PubMed literature watch, citation reports, and new-PMID alerts Public-sector bid-alert queue, award watch, rebid signals Discover current article URLs by news query
What it reads Official NCBI ESearch and ESummary JSON TED, SAM.gov, and Grants.gov notice APIs Google News RSS
Primary input PubMed searchTerms or known pmids Jurisdictions, keywords, CPV/NAICS codes News queries (max 50)
Primary output Citation metadata, literature-update reports, citation exports, new-publication alerts Normalized tenders/grants, alerts, optional procurement report/export Headline, article URL, publisher, timestamps, snippet
Full text / HTML cleanup No. Citation metadata only; not abstracts or full text Not a page extractor No. Discovery only; not article bodies

This independent Actor is not affiliated with, sponsored by, or endorsed by NCBI or PubMed. PubMed and NCBI are public data sources used by the Actor; they do not sponsor, approve, or endorse it.

Use cases

How is PubMed Literature Watch different from TED, SAM.gov & Grants Bid Alerts and Google News Scraper?

This Actor queries official NCBI E-utilities (ESearch and ESummary JSON). It returns PubMed citation metadata, literature-update reports, systematic-review citation exports, and new-publication alerts for saved queries. It does not scrape a browser UI and it does not log in. TED, SAM.gov & Grants Bid Alerts Scraper is a public-sector procurement queue: TED tenders, SAM.gov opportunities, and Grants.gov notices. Google News Scraper is query-based Google News RSS discovery of headlines and article URLs. This page is biomedical literature watch on PubMed, not procurement and not news RSS.

What is the best first run?

The published README splits two first-run paths. For a low-cost new-publication signal, use a bounded monitor. The baseline and unchanged poll emit zero rows and zero charges:

{
  "searchTerms": ["machine learning healthcare"],
  "sort": "most_recent",
  "maxResultsPerQuery": 10,
  "maxArticles": 10,
  "runMode": "monitor",
  "monitorKey": "starter-machine-learning-healthcare",
  "initialRunMode": "emit_backfill",
  "emitRawRows": false,
  "generateReport": false,
  "generateExport": false,
  "maxChargeUsd": 0.30,
  "dryRun": false
}

This starter is a discovery signal, not a literature report or medical recommendation. It can emit up to ten new-publication-alert rows at the configured $0.03 each, for a maximum planned value charge of $0.30.

Use this separate path when you need a grouped literature update and citation-metadata export. With the configured prices, a non-empty report plus export plans $7.00 total ($2.00 + $5.00). A query with no matches emits no value rows and creates no value-event charge. Do not force the $2.00 report event under a $1.00 cap:

{
  "searchTerms": ["cancer immunotherapy"],
  "fromDate": "2026",
  "sort": "most_recent",
  "maxResultsPerQuery": 25,
  "maxArticles": 25,
  "runMode": "audit",
  "emitRawRows": false,
  "generateReport": true,
  "generateExport": true,
  "maxChargeUsd": 7,
  "dryRun": false
}

Schema prefill for searchTerms is cancer immunotherapy. Live schema additionalProperties is false; send only the published field names below.

Open PubMed Literature Watch → for a cheap signal use monitor + report/export OFF; for a literature update keep audit + report/export ON under a $7 cap. Citation metadata only — not full text.

How do audit, monitor, and extract modes work?

Live schema runMode default is audit. The schema description: report-first audit generates the current report; monitor emits new PMIDs; extract is the compatibility path for raw metadata rows. Default generateReport and generateExport are true. emitRawRows default is false. Leave raw rows off for the report/export value path.

Published README input examples:

Current literature report:

{
  "searchTerms": ["GLP-1 obesity"],
  "fromDate": "2025",
  "maxResultsPerQuery": 25,
  "maxArticles": 50,
  "runMode": "audit",
  "emitRawRows": false,
  "generateReport": true,
  "dryRun": false
}

PMID metadata export. Use raw extraction only when an integration explicitly needs article, author, or journal metadata rows. Raw rows use the configured $0.004 dataset event; they are not the report-first value path:

{
  "pmids": ["39763750", "39763800"],
  "runMode": "extract",
  "emitRawRows": true,
  "generateReport": false,
  "generateExport": true,
  "dryRun": false
}

New-publication monitor. The first baseline_only monitor emits no rows. Later runs emit only newly observed PMIDs; an unchanged run writes zero rows and creates zero charges:

{
  "searchTerms": ["CAR-T solid tumor"],
  "sort": "most_recent",
  "maxArticles": 100,
  "runMode": "monitor",
  "monitorKey": "car-t-solid-tumor-watch",
  "initialRunMode": "baseline_only",
  "emitRawRows": false,
  "generateReport": true,
  "generateExport": false,
  "dryRun": false
}
Field Type Default Notes
searchTerms string[] [] PubMed query strings passed to NCBI ESearch. Schema prefill: cancer immunotherapy.
pmids string[] [] Known PubMed IDs to summarize directly.
fromDate string empty Optional publication date lower bound accepted by PubMed.
toDate string empty Optional publication date upper bound accepted by PubMed.
sort string most_recent relevance, pub_date, or most_recent.
maxResultsPerQuery integer 25 Maximum PMIDs fetched from each query; 1–200.
maxArticles integer 25 Maximum deduplicated PubMed records processed in one bounded run; 1–500.
runMode string audit audit, monitor, or extract.
monitorKey string empty Stable private state namespace for a scheduled PubMed query.
initialRunMode string baseline_only baseline_only or emit_backfill.
emitRawRows boolean false Optional compatibility output. Raw citation, author, and journal rows use the dataset event.
generateReport boolean true One non-empty query-level literature update report. No matches means no report event and no report charge.
generateExport boolean true One citation-metadata export for screening workflows. No matches means no export event and no export charge.
email string empty Optional registered developer contact sent to NCBI.
apiKey string — Optional free NCBI key for up to 10 requests/second. Keyless mode is supported at 3 requests/second.
tool string pubmed-research-intelligence Tool identifier sent to NCBI E-utilities.
timeoutMs integer 20000 Maximum time for each NCBI request; 1000–60000.
delivery string dataset dataset or webhook. Webhook also POSTs the billed row payload to an HTTPS URL.
webhookUrl string — Optional public HTTPS destination when delivery is webhook.
maxChargeUsd number 7 Optional Actor-side spending cap; 0–1000. Default report plus export plans up to $7.00. If a cap is too low, delivery stops before any row or state is committed.
dryRun boolean false Return illustrative rows without NCBI requests or billing.

The live schema has no required array. additionalProperties is false. Published Store example run input (schema-valid; omitted fields take schema defaults, including runMode audit, generateReport true, and generateExport true):

{
  "searchTerms": [
    "cancer immunotherapy",
    "GLP-1 obesity"
  ],
  "pmids": [],
  "fromDate": "2024",
  "toDate": "",
  "sort": "relevance",
  "maxResultsPerQuery": 10,
  "maxArticles": 20,
  "email": "",
  "apiKey": "",
  "tool": "pubmed-research-intelligence",
  "timeoutMs": 20000,
  "delivery": "dataset",
  "webhookUrl": "",
  "dryRun": false
}

What does a result contain?

The published README sample is a literature_update_report row. It is a README illustration, not a live coverage guarantee:

{
  "rowType": "literature_update_report",
  "billingEventName": "literature-update-report",
  "query": {
    "searchTerms": ["cancer immunotherapy"],
    "fromDate": "2026"
  },
  "articleCount": 12,
  "pmids": ["39763750", "39763800"],
  "topJournals": [{"value": "Cancer Research", "count": 3}],
  "summary": "12 PubMed publications matched this literature update.",
  "sourceUrl": "https://pubmed.ncbi.nlm.nih.gov/"
}

An export uses the same matched citations in a reusable citations array. A monitor run adds one new-publication-alert for each newly observed PMID; the first baseline_only run and an unchanged run emit zero rows and zero charges. Raw extract rows are citation, author, or journal metadata billed as apify-default-dataset-item.

Does this Actor return full text or abstracts?

No. The published README Source and limits section states that this Actor exports citation metadata, not article full text or abstracts. It uses official NCBI ESearch and ESummary JSON endpoints; no browser scraping or login is required. An optional free NCBI API key increases the documented request allowance; keyless mode is paced conservatively. PubMed records can be incomplete or updated after publication.

Outputs support literature research and screening workflows. They are not medical advice, clinical efficacy claims, or a substitute for reviewing the source publication. Keep NCBI attribution and source URLs in downstream reports.

How is PubMed Literature Watch & Research Report priced?

Billing is pay per event. The live Store card is from $4.00 / 1,000 pubmed metadata rows. Current PPE:

Event Price Emitted when
apify-default-dataset-item (PubMed metadata row) $0.004 One delivered citation, author, or journal metadata row
new-publication-alert $0.03 One new source-linked publication matched by a saved monitor
literature-update-report $2.00 One generated query-level literature update report
systematic-review-export $5.00 One citation metadata export for screening or reference workflows

There is no start charge. The default input is report-first (audit, raw rows off, report and export on). Use maxChargeUsd to cap the complete planned delivery: if the cap is too low, no rows are delivered and monitor state is not committed. A monitor baseline-only run and an unchanged monitor poll both cost $0.00. The live pricing tab lists Platform usage as included.

from $4.00 per 1,000 pubmed metadata rows; $0.03 per new-publication alert; $2.00 per literature-update report; $5.00 per systematic-review export

See PubMed Literature Watch & Research Report pricing on Apify

Limits to keep in mind

Published README next-report landings:

Open PubMed Literature Watch → for a cheap signal use monitor + report/export OFF; for a literature update keep audit + report/export ON under a $7 cap. Citation metadata only — not full text.

Related pages