How we source & verify

RecapFeed is an independent aggregator, not a newsroom. That only works if you can trust where the information comes from — so here's exactly how the site gathers, filters, verifies and publishes everything you see. Every headline, score and image belongs to its original source and links straight back to it; we add the organising, not the reporting.

Where the data comes from

• News, reviews & videos — RSS feeds published by gaming outlets (IGN, Eurogamer, GameSpot, Push Square, Nintendo Life, Polygon, VGC and many more) plus YouTube channels for trailers and review videos. Each item links to the outlet that published it.
• Release dates, developer, publisher & genre — cross-referenced across Steam, RAWG, Wikidata and Wikipedia's per-year release lists, so no single source is taken on its own.
• Cover art — Steam and SteamGridDB, matched to each game by exact title so the wrong art can't attach.

How we verify release dates

Dates are the thing people plan around, so they get the most scrutiny. When sources disagree, the pipeline arbitrates: a firmer precision wins (a confirmed day beats a vague "Q4"), and a current or future date beats a stale one. Every game's page shows which source its date came from and when the page was last refreshed, so you can judge it yourself. Release dates shift constantly — always confirm with the publisher before planning around one, and if you spot a wrong date, tell us and it gets fixed.

How we filter

Not everything in an outlet's feed is gaming news. Before anything is published, high-precision filters drop:

• Commerce — affiliate deals, price drops and "prebuilt PC for $X" posts.
• Off-topic entertainment — film, TV and toy coverage that leaks from outlets' all-content feeds.
• Service content — walkthroughs, unlock guides, tier lists and code lists. Useful, but not news.

These filters are deliberately conservative: they're tuned to avoid false positives, so we'd rather let a borderline item through for a second look than wrongly drop real news.

The curation pass — and where AI fits

After the mechanical fetch and before anything goes live, a curation pass (AI-assisted, using Claude) runs on top of the filters. It does three things: drops off-topic items the filters miss, picks the home page's featured stories by importance, and writes the daily Today brief. We're upfront about this — the Today brief is clearly labelled AI BRIEF. Crucially, the model works only from the wording of headlines published by named outlets; it does not invent facts, numbers, dates or scores, and nothing is published until it has been vetted. If the AI step is unavailable, the site simply falls back to the filtered feed — it never blocks a refresh.

Freshness

RecapFeed rebuilds from its sources roughly every two hours — a refresh is a full rebuild, so the page you load is always freshly baked, not cached stale. The exact time of the last refresh is shown in the footer of every page.

What we deliberately don't do

• We don't publish critic scores from an API or invent our own — where a review score isn't in the source wording, we don't show one.
• We don't rehost anyone's images or articles — they load from their original CDNs and link back to the source.
• We don't run affiliate or deals content — RecapFeed makes no money from what it links to.

Independence & corrections

RecapFeed isn't affiliated with any outlet it aggregates. Headlines, scores and images belong to their original sources, credited and linked on every item. Found something wrong — a bad date, mismatched art, an item that shouldn't be here? Reach Ben via benford.co.nz or the corrections note, and it'll get fixed. More about who's behind the site on the About page.