Feed Delta — RSS, Atom and JSON Feed Updates

Get new and updated feed entries as clean JSON, CSV or Excel data. Feed Delta compares each entry with a previous snapshot, so a workflow can skip unchanged articles and still detect edits to older entries.

Use it for newsletter research, feed-to-database imports, release-note monitoring and collecting updates for an AI workflow. It reads feed content; it does not crawl the full article pages or generate summaries.

Try it

1. Keep the prefilled GitHub blog example, or enter 1–5 public HTTPS feed URLs.
2. Click Start. Open Output to view or download the entries.

No proxy, browser or additional API key is needed. The GitHub blog is an example source, with no affiliation.

Get only changes on later runs

The first run treats all valid entries as new. For incremental processing:

1. Open the run's key-value store and copy NEXT_STATE.
2. On the next run, expand Incremental processing and use that object as Previous state.

In an automated workflow, retrieve NEXT_STATE from the run's default key-value store, save it after your downstream delivery succeeds, and pass it as previousState on the next run. Each workflow owns its snapshot, so separate workflows do not overwrite one another's progress. Running on a schedule alone does not transfer this state.

{
  "feedUrls": ["https://github.blog/feed/"],
  "maxResults": 100,
  "previousState": {"version": 1, "seen": {}}
}

An empty previousState starts a new baseline. The full snapshot contains fingerprints, not article text.

What you receive

The default dataset contains only new or changed entries:

Field	Meaning
change	new or updated
id, key	Feed identity and a stable key scoped to the feed URL
title, url, content	Content supplied by the feed
source	Original feed URL
published, updated	Feed-provided timestamps, when available
content_hash	Fingerprint for downstream deduplication

Download the dataset as JSON, CSV or Excel using Apify's export controls. The key-value store also contains:

• OUTPUT: counts, source formats, errors and any results not emitted because a limit was reached.
• NEXT_STATE: the snapshot to use on the next run.

A successful run with zero dataset rows means there were no changes, provided OUTPUT.status is completed.

Pricing

• $1 per 1,000 new or updated entries delivered ($0.001 each).
• $0.00005 per start at the default 256 MB memory. Apify charges one start event per GB, with a minimum of one.
• Unchanged entries have no result charge. A no-change run still has its start charge.

For example, 100 delivered entries at the default memory cost $0.10005. The platform's current pricing panel is authoritative. Set a maximum cost per run to control spending; the minimum allowed limit is $0.002. The configured event prices include ordinary platform usage.

When a result or spending limit stops delivery, NEXT_STATE includes only delivered changes. Pass that state to retrieve the remaining changes later. A platform interruption can happen before the final summary is written; inspect the run status and available records before advancing your downstream checkpoint.

Limits and behavior

• RSS 2.0, Atom, RDF/RSS and JSON Feed; UTF-8; up to 5 MB per feed.
• 1–5 public HTTPS feeds on port 443, up to 1,000 changed entries per run and 50,000 remembered identities.
• Entries need a stable ID or link. Identity is scoped to the exact feed URL, including its query string.
• Missing entries are not deletions: feeds typically retain only recent content.
• Source failures appear in OUTPUT. Good sources can still produce results, while total source failure fails the run. Failed sources do not remove previously saved fingerprints.
• No login, browser rendering, proxies or XML DTD processing.
• A crash between delivery and checkpointing can replay an entry. Downstream consumers should deduplicate using key plus content_hash; exactly-once delivery and charging are not guaranteed.

For a problem, open the Actor's Issues tab with the run ID, feed format and a short description. Do not include credentials or private feed content.