How we check

1. The journal’s own page first. A fee, submission or author-information page on the journal’s or publisher’s site, read on a stated date and quoted in the data file.
2. The publisher’s price list second. Where a publisher publishes a list covering many journals (Elsevier, EDP Sciences, Springer Nature’s pilot), we use the row for the journal and say so.
3. A platform rule only when it is explicit. OpenEdition Journals admits only journals that charge no submission or publication fees, so its journals are listed as free with that rule as the source.
4. An archived copy when the live page blocks automated reading. The row then names the archive date. Several large publishers block scripts; we do not bypass the block.

• A figure that we could only find in a directory, a search-engine summary, a blog or an email. The row says “unclear” instead, and the reader is told not to read that as free.
• A converted currency. Each amount stays in the currency the publisher states.
• A country waiver inferred from an income list a publisher does not itself use. Where a publisher does not publish a country list, we say so instead of inferring one.

Reductions apply to fully open access journals unless the publisher says otherwise; hybrid titles are noted separately. Eligibility is read from the corresponding author’s affiliation at submission, because that is what the publishers’ systems read. Where two publishers disagree about a country, we show both figures rather than an average.

Research4Life Group A and B lists are taken from research4life.org/access/eligibility, the only public source (confirmed by Research4Life, 12 October 2026). Research4Life keeps no changelog; upcoming moves are announced as footnotes on that page a few months ahead. We check the footnotes monthly.

When we ask a publisher about its page, its answer is recorded on Publisher responses with the date. An emailed answer is shown as the publisher’s statement; the figure still comes from its page, and where the two differ we show both.

Every row carries the date we read the page. Country lists and large publishers are re-read when we have reason to think they changed; on 11 October 2026 the Wiley list was re-read in full and one country had been removed since August. Changes to published figures go in the corrections log.