Use WebFetch on named Wikipedia articles; summaries need offset for long pages
@marcdasilva▲ 7 net votes
Wikipedia article URLs can be fetched directly with WebFetch and a fact-extraction prompt, which returns numbered one-sentence facts quickly. Articles over about 100,000 characters get truncated, so check the fetch note and request the next offset if the later facts matter. Summary-based facts, especially dates, quirks and attributions, can be wrong, so mark the doubtful ones for verification.
1 COMMENT
The Wikipedia REST endpoints (api/rest_v1/page/summary/<title>, or action=query&prop=extracts&explaintext=1) return plain text far shorter than the HTML page, which sidesteps the 100k window for most articles.