INSPIRE-HEP Literature Monitor
Pricing
from $20.00 / 1,000 query checkeds
INSPIRE-HEP Literature Monitor
New INSPIRE-HEP records for a saved query since the last run. Stores title, record id, dates, citation count, arXiv id, and DOI. Drops email addresses. Not affiliated with INSPIRE.
Pricing
from $20.00 / 1,000 query checkeds
Rating
0.0
(0)
Developer
Muhammet Mustafa Dincer
Maintained by CommunityActor stats
0
Bookmarked
1
Total users
1
Monthly active users
2 days ago
Last modified
Categories
Features
Share
Watches a saved INSPIRE-HEP literature query and reports control numbers that were not in the previous run. The sample query is control_number:2181837 or arxiv:hep-th/9711200. On 7 Nov 2026 that query returned two records from https://inspirehep.net/api/literature: control number 2181837, arXiv 2211.07478, titled "Study of $HZZ$ anomalous couplings by angular differential cross sections", and control number 451647, arXiv hep-th/9711200, with DOIs 10.4310/ATMP.1998.v2.n2.a1 and 10.1023/A:1026654312961. No customer API key is required. The request goes only to the public literature endpoint.
This Actor is not affiliated with INSPIRE, INSPIRE-HEP, or CERN.
The stored fields are the title, the INSPIRE control number, the earliest date, the preprint date, the record created and updated times, the citation count, the arXiv id, and the DOI. Author names, affiliations, abstracts, and email addresses are not stored. INSPIRE's terms allow reuse of HEP collection metadata under the CC0 waiver and do not permit downloading email addresses. An email address in a kept field is removed before the row is written. The Actor does not reconstruct an article from snippets, and it does not call /api/accounts, /api/submissions, or /api/literature/import.
Store actors that read INSPIRE today return a search dump. They do not keep a baseline, and they do not report which control numbers appeared after the last successful read. parseforge/inspire-hep-scraper lists a result event at $0.005 on the free tier. automation-lab/inspire-hep-literature-search lists a paper event at about $0.0045 on the free tier. johnvc/google-scholar-api lists a search query at $0.015 and is the research-paper neighbor with current demand. This Actor is the scheduled check: which control numbers in this saved query were not saved last time?
Quick start
- Keep the sample query.
- Leave Only new records off. The sample prefill does that on purpose.
- Click Start.
The first successful read writes one row per record with status baseline_saved. A second run of the same sample still writes those rows, with status unchanged, because the prefill does not hide a control number that was already saved. The rows are the current INSPIRE metadata. They are not invented papers.
Turn Only new records on when you schedule the Actor. Omitting the field is the same as turning it on. After a baseline exists, an unchanged control number stays out of the dataset and still appears in the run summary. A first successful read is always written. A control number that was not saved is new_record. An explicit onlyNewRecords: true keeps the quiet schedule. An explicit false always writes the current window.
Use a different Watchlist name to start a fresh baseline. Baselines live in the named key-value store inspire-hep-literature-monitor-state. The snapshot stores control numbers only, not titles and not addresses. Under limited permissions, if that store cannot be opened, the run uses the default store, finishes, and treats the look as a first baseline.
The read is one request, sorted by most recent, limited to Max records. The sample asks for 5, and the sample query currently matches 2. A record that falls outside that window is not reported as removed. Up to 5 queries can be checked in one run. Default memory is 512 MB, so the platform start event is one unit. A normal response returns in a few seconds.
The run does not call INSPIRE when the spending limit cannot pay for query-checked. An empty result, an API error, a network error, and an invalid query are not charged, and an empty or failed read does not erase a baseline already saved. A new control number that does not fit in the spending limit is not written to the dataset and is not added to the baseline. charged is true only when Actor.charge() succeeded for that row. An ignored charge, including an owner run with no price, still saves the records and leaves charged false. A query check that was already billed stays billed when a later new record does not fit.
Pricing
The prices in pricing-ppe.json are a proposal. They are not applied on the platform from this repository.
| Event | When | Free tier |
|---|---|---|
apify-actor-start | The run starts. At 512 MB this is one event. | $0.0005 |
query-checked | One saved query whose first page included at least one record, including a later run with no new control numbers. | $0.015 |
new-record | Extra charge when a control number on that page was not in the saved baseline. The first baseline is not this event. | $0.005 |
query-checked copies the published tiers for the johnvc/google-scholar-api event "Search API query executed": free $0.015, Bronze $0.01485, Silver $0.014701, and $0.014554 from Gold through Diamond. new-record copies the published tiers for the parseforge/inspire-hep-scraper event "result": free $0.005, Bronze $0.00484, Silver $0.00468, Gold $0.00452, Platinum $0.004068, Diamond $0.003616. apify-actor-start is $0.0005, the start price used by the published change monitors in this account. Google Scholar's own start event is $0.00001 beside a separate setup fee, and both OpenAlex and automation-lab/inspire-hep-literature-search list a $0.005 start. This proposal does not add a setup event, because the code does not charge one.
Each later tier is lower than or equal to the tier before it. The platform margin in the pricing file is 20 percent. pushData is called without an event name, so a row is not billed twice.
Input
- Saved queries. INSPIRE literature search syntax, one per line. The sample is a single line that matches two stable records.
- Max records. An integer from 1 to 25. The sample and the schema default are 5.
- Only new records. Sample prefill
false. Schema defaulttruewhen the field is omitted. - Watchlist name. Default
default. Letters, numbers, spaces, dots, underscores, and hyphens.
There is no API key field. Do not put an email address in the query. Addresses are stripped, and a query that is only an address is rejected.
Output
RUN_SUMMARY counts baselines, unchanged records, new records, empty queries, withheld records, and warnings. The run fails when every requested query fails to produce a baseline, an unchanged read, a new record, or an empty result, and when every line was invalid. A run that stops because of the spending limit still succeeds and leaves a not_fetched row.
License
Apache-2.0. INSPIRE is a high-energy physics literature service. This Actor is not affiliated with INSPIRE. The sample records are a public metadata read, not an endorsement.