RSS Feed Reader – RSS & Atom Feeds to JSON

Give it feed URLs and get each feed back as one JSON row with its entries: title, link, summary, content and dates. Use it to pull RSS into a database, an agent or a workflow without writing an XML parser. To get only what is new since the last run, use the RSS Feed Monitor Actor instead.

Read supplied public HTTP(S) RSS 2.0, RSS 1.0 or Atom feeds. One dataset row per processed distinct feed, containing up to 100 snapshot entries. No feed discovery, article fetching, state, delta tracking or broad brand coverage.

{"urls": ["https://www.nasa.gov/feed/"]}

Output and limits

Each row contains normalized url, finalUrl, httpStatus, format (rss2/rss1/atom, or null), entries, truncated, elapsedMs and reasonCode. Each entry has nullable id (RSS 1.0: the item's rdf:about), title, link, summary, published (RSS 1.0: dc:date), updated (Atom) and content (Atom <content> or RSS <content:encoded>, the body as supplied by the feed, with Atom XHTML child markup preserved). Only supplied values are returned; no invented IDs or parsed/inferred dates. RSS guid remains opaque even if not a permalink; text fields, including guid, have surrounding whitespace trimmed. Atom updated is returned as updated, never as a publication date. Relative entry links resolve against the final feed URL (XML base overrides are not used). Only HTTP(S) links without credentials are emitted; they are not fetched or checked.

Titles, summaries and content may contain untrusted HTML/XHTML. Treat output as data; sanitize before rendering.

• 20 URLs/run, before dedup; canonical repeats including fragments/default ports read once. Paths and queries stay distinct; different invalid inputs stay separate.
• 100 entries/feed, and at most 5 MB of entry text per row (long podcast show notes can reach it first); the first entries in feed order are kept. truncated: true means additional entries were parsed but omitted. This flag covers entries only, not unprocessed feeds. No claim about entries the source did not publish.
• 32 MiB/feed (after decompressing gzip/deflate/br, which some servers send even when asked not to), enough for large podcast feeds; 15 seconds total/feed including DNS, 5 redirects. No retries. UTF-8, BOM-marked UTF-16, and any standard encoding the XML declaration names (ISO-8859-1, windows-1252, Shift_JIS, EUC-KR…) are decoded; an unknown encoding fails free. XML nesting capped at 64.
• All DNS addresses must be public at every hop; transport pins a checked IP and verifies TLS. Public hostnames and ports 80/443 only; no credentials or IP-literal URLs.
• DTD/custom entities rejected; ordinary predefined XML escapes and numeric character references supported. Malformed XML and nonfeeds are free.

Charge caps stop batches early. Unprocessed feeds have no dataset row and no charge; an absent row is not an empty feed. To identify missing feeds, compare your submitted URLs after canonical normalization (trim whitespace, normalize hostname case/default ports and remove fragments; keep paths and queries distinct) with persisted normalized url fields, not finalUrl. Raise the charge cap before resubmitting only the missing URLs.

Billing

$0.002 per answered distinct feed, not per entry. A parsed empty feed is an answer and bills. A truncated but successfully parsed snapshot is an answer and bills. All failures below are free.

Reason code	Meaning	Charged
read	Successfully parsed RSS 2.0, RSS 1.0 or Atom snapshot, including empty	yes
invalid_url	Unsupported/malformed URL	no
unsafe_target	DNS contains non-public addresses	no
blocked	HTTP 401/403/429	no
http_error	HTTP 206 partial response or other non-2xx status, including 404/410	no
unreachable	DNS, transport, TLS, encoding or deadline failure	no
invalid_redirect	Missing, malformed or disallowed redirect	no
redirect_loop	Repeated redirect URL	no
redirect_limit	More than five redirects	no
too_large	Response exceeds 32 MiB	no
invalid_xml	Malformed/unsupported XML, DTD/entities or excessive nesting	no
not_feed	Parsed document is not RSS 2.0, RSS 1.0 or Atom	no

A failed or empty snapshot is not proof a publisher has no articles elsewhere. Feed contents are whatever the publisher returned at request time, not a historical archive.

Source rights

You supply the feed and are responsible for permitted access, attribution and reuse. Public feed access does not grant all redistribution rights. This reader does not bypass authentication, bot walls or access restrictions.

Default/local samples are provided by NASA, listed on NASA RSS feeds. NASA content guidelines permit factual informational use without implied endorsement, require acknowledgment, and retain third-party copyright, logo and personality restrictions. NASA is the source, not an endorser. Samples read text metadata only; no images or linked articles are fetched. This is not legal approval for arbitrary customer feeds.