RSS & Atom Incremental Feed API

Read public RSS and Atom feeds with stable item keys, deduplication and source timestamps for incremental workflows.

What you can do

Build a news inbox or a feed-to-dataset integration without processing the same item twice.

Quick start

{
  "feedUrls": [
    "https://blog.python.org/feeds/posts/default?alt=rss"
  ],
  "maxItems": 10
}

Run the Actor and download the default dataset as JSON, CSV or Excel. Inspect the OUTPUT key-value-store record before treating the run as complete. Live example sources can change or become unavailable.

Behavior and limitations

RSS 2.0, RDF/RSS and Atom are supported. Pass itemKey values from the previous dataset as seenKeys. Identity is scoped to the input feed URL, so IDs reused by different publishers do not collide. Input feed URLs should remain stable between runs. Publisher date strings are preserved as publishedAtRaw; they are not guessed or converted. Feed entries are returned in source order. Summaries are plain text capped at 10,000 characters. This is a snapshot of currently exposed feed entries, not a historical archive. A changed item with the same ID is filtered if its key is already in seenKeys.

Every result preserves its source URL and an observedAt UTC timestamp. Network or parsing errors are listed by source in OUTPUT.errors; other successful sources still return results. OUTPUT.warnings, complete and status distinguish a full observation from partial output. If every requested source fails, the run fails rather than silently returning an empty success.

Pricing

USD 0.8 per 1,000 successful new feed item rows ($0.0008 per row), plus a $0.005 Actor-start event for the default 256 MB run. No result event is charged for a failed source. The Actor-start event can still apply to a failed run. Higher memory can increase start-event quantity according to Apify's rules. This description must match the currently effective Store pricing; check the pricing tab before running. Apify charges only as configured there.

Resource and access limits

At most 1,000 output rows; public ports 80/443 only; five redirects; 2 MiB decoded body per request; approximately 100 seconds for fetching. Requests are sequential with no retries, CAPTCHA bypass, credentials, residential proxies or paid upstream APIs. robots.txt is respected; missing robots.txt (404/410) allows fetching, while blocked or unavailable policies produce errors. Public DNS addresses are validated and pinned for each connection. Use only sources you are permitted to access and respect source terms. Private networks and URLs with embedded credentials are rejected.

API and automation

Call POST https://api.apify.com/v2/actors/soilair~rss-atom-incremental-feed-api/runs with the example JSON and your own Apify authorization header. Poll the returned run ID, then fetch /v2/datasets/{defaultDatasetId}/items. Schedule subsequent runs in Apify Console; pass earlier keys/hashes explicitly where supported. The Actor never schedules jobs itself.

Data handling and support

Input and results remain in your Apify run's storage under the platform's retention settings. The Actor sends requests only to the input sources, their public redirects and robots.txt. No separate analytics or customer-data endpoint is used. Do not put secrets or private customer URLs in the input. Developed with AI assistance and tested with deterministic fixtures plus public-source smoke checks; this does not guarantee all storefronts or websites will work. For support, send the run ID, source URL and expected behavior via the Apify issue tab without credentials.