RSS News Dataset & Deduplicator
Pricing
Pay per usage
RSS News Dataset & Deduplicator
Aggregate public RSS and Atom feeds into a normalized, cross-source deduplicated article dataset with failure-safe NEW/SEEN snapshots.
Pricing
Pay per usage
Rating
0.0
(0)
Developer
X L
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
7 days ago
Last modified
Categories
Share
Aggregate public RSS, Atom, and RDF feeds into one normalized article dataset. The Actor collapses the same canonical article across feeds, strips common tracking parameters, filters by keyword or publication window, and stores a snapshot for NEW versus SEEN scheduled-run workflows.
What it returns
Each article row includes the normalized article URL, title, publication date, author, categories, summary text, bounded content text, source feeds, duplicate count, matched keywords, and incremental status.
Feed failures stay visible in SUMMARY.feedErrors while successful feeds continue. REPORT provides a compact Markdown run summary and article preview.
Inputs
feedUrls: up to 20 public RSS, Atom, or RDF URLsmaxItemsPerFeed: parsed-entry cap per feedmaxItems: final dataset cap after filtering and deduplicationmaxRetainedBytes: run-level normalized-data memory budgetincludeKeywords/excludeKeywords: case-insensitive article filterspublishedAfterDays: optional publication window; dated filtering excludes entries without a parseable datestripTracking: remove common tracking query parameters before deduplicationsnapshotKeyandsaveSnapshot: incrementalNEW/SEENstate
Guardrails
- Public HTTP(S) feeds only; private, loopback, reserved, and local network addresses are rejected.
- DNS results are pinned for each request so a hostname cannot be re-resolved to a private address after validation.
- Feed URL length, redirects, request time, feed bytes, feed count, concurrency, retained normalized data, and output rows are capped.
- XML parsing runs in an isolated worker with time and heap limits; item normalization is capped before article objects are built.
- Failed feeds retain their prior per-feed snapshot state. A run where every feed fails exits as failed without replacing the snapshot.
- XML entity processing is disabled; output text is cleaned after parsing.
- No browser, proxy, API key, or external AI service is required.
Local validation
npm testnpm run smoke
The first Store release should remain free while organic usage is measured. A future PPE boundary is article-emitted, charged only for visible normalized article rows, never failed feeds or entries removed before output.